U.S. Army Medical Department, Office of Medical History
Skip Navigation, go to content







AMEDD MEDAL OF HONOR RECIPIENTS External Link, Opens in New Window






Chapter XIV


Chapter XIV


The investigations herewith presented were carried out in 1917-18 in the laboratories of Yale University, New Haven, Conn., and American University, Washington, D. C., under the direction of the medical science section of the Chemical Warfare Service.a These investigations had their inception in studies on the general problems of war gases, begun in 1917 by the Bureau of Mines of the Department of the Interior, and later transferred to the newly organized Chemical Warfare Service. The scope of the work widened so rapidly that a division into the separate fields of physiology, pathology, and therapy was deemed advisable. The studies in pathology, with which this chapter is concerned, were under the direction of Maj. M. C. Winternitz, M. C., professor of pathology, Yale University.

At the Yale station the work was done chiefly in the Brady Laboratory of Pathology and Bacteriology, in close collaboration with the department of physiological chemistry, which had charge of investigations on physiology and therapy. At American University there was similar cooperation between the pharmacologists and pathologists, thus making possible at each institution the investigation of a large amount of material.

In order to avoid duplication of effort the two stations took up the study of different groups of gases. At the Yale station chlorine, chloropicrin, and phosgene were investigated; at American University arsine, organic arsenicals, superpalite, cyanogen compounds, and mustard gas were studied. Special problems relating to the effects of these gases were similarly assigned. Unfortunately, from the scientific standpoint, the work ended rather abruptly on December 1, 1918, leaving unfinished the solution of a number of problems.


Chlorine, first of the toxic gases used in the World War, was among the first to be subjected to experimental study by the Medical Division of the Chemical Warfare Service. In the course of the pharmacologic and therapeutic studies upon this gas, carried on in Washington and New Haven, the results of which are discussed elsewhere in this volume, animals in large numbers were gassed, thus providing abundant material for pathologic investigation.

In these studies the dog was the experimental animal of choice, although, as with other gases investigated, various other animals (guinea pigs, rats, mice, rabbits, cats, monkeys, and goatsb) were used by way of comparison. The reason for using the dog is that, in the study of the respiratory irritant

a The data in this chapter are based on reports prepared under the direction of Maj. M. C. Winternitz, M. C., and published as: "Collected Studies on the Pathology of War Gas Poisoning," New Haven, Conn., Yale University Press. 1920.
b The goat as an experimental animal was preferred by some British investigators, partly because of the larger size and distinct lobulation of the lungs, and also because of the resistance of this animal to infection, which permitted a better study of the uncomplicated gas injury. In the conduct of the investigations here discussed the very large size of the lungs was not believed to be an advantage; and the fact that in their relative immunity to respiratory infection goats differ markedly from man was believed to he a reason against rather than for their use, except in so far as the absence of infection permitted the study of uncomplicated gas injury.


gases, to which group the majority of war gases belong, this animal has certain well-recognized advantages: (1) The lungs and other organs are sufficiently large to make gross examination easy. (2) Anatomically the respiratory tract resembles closely that of man. (3) The conditions of pulmonary infection and the reaction of the lungs to bacterial and other injuries are much the same as in human beings.

The accompanying diagram (Fig. 46) shows very well the topography of the dog's respiratory tract. The trachea is considerably longer than the human organ, but it branches in similar manner, as shown in the diagram. As in man, the right lung has three lobes and the left two. The divisions are more complete, however, and the left upper lobe has a deep fissure, nearly dividing it into two equal parts. There is a sixth, or caudate, lobe which, as the diagram shows, may be considered as belonging to the right lung, although its bronchus is given off close to the tracheal bifurcation, and its position behind the pericardium is rather neutral. c

The relations of the bronchial tree are shown very well in Figure 47. It is a fact of interest and importance, as the succeeding report will show, that the respiratory irritant gases exhibited striking differences in their action on the several segments of the air passages: Mustard gas, for example, damaged chiefly the first portion-that is, trachea and large bronchi; others, such as phosgene and superpalite, affected only the distal portion; while chlorine, the gas under consideration, injured the entire tract.

FIG. 46.- Diagram of dog's respiratory tract viewed from behind. The interlobular fissures and the accessory caudate lobe are well shown

The material upon which the present report is based includes 326 dogs that succumbed or were killed after exposure to chlorine in gassing chambers. The time of exposure was in most cases 30 minutes; the concentration varied between 600 and 1,100 parts per million.

c In considering the pathological anatomy and physiology of the lung of the dog or other quadruped, it should be kept in mind that the anterior portion of the organ and not the so-called lower or diaphragmatic portion, as in man, is the most dependent in the normal posture of the animal. The significance of this fact will he discussed in connection with the localization of certain of the lesions following gassing.



While death following severe exposures often occurred within a few minutes and was generally not delayed beyond a few hours, the exposed animal might survive the acute period, only to succumb after days or even weeks. As might be expected, the pathological changes found at autopsy varied markedly with the period of survival, and it will make for clearness of description if in the discussion of these changes, the cases studied are divided into three groups: (1) Acute deaths; (2) delayed deaths; (3) chronic or recovered cases.

FIG. 47.- Diagram of bronchial tree of dog, showing bronchi of flrst, second, and third order (A, B, C) and infundibula (D)

In Table 49 the animals are grouped according to the duration of life after gassing. It is seen that a majority of the fatalities occurred in the first 24 hours, 172 out of 270, or 64 percent. When figures are plotted (Chart XXIV) it is also seen that the death curve starts with a sharp rise, which is followed by an equally sharp fall. There is a short secondary rise on the fourth day. This second rise, it is believed, is to be attributed to the development of pulmonary infection, a phenomenon that will be emphasized in a succeeding paragraph.

TABLE 49.- Dogs gassed with chlorine




The gross changes seen at autopsy, when death occurred in the first 24 hours after gassing, were remarkably uniform. Such differences as were noted were largely those of degree. In general it may be said that the changes, particularly in the respiratory system, became progressively more marked as death was delayed, up to 24 hours, when the acute, uncomplicated lesions had reached their maximum.

As a rule, the eyes were reddened, showing an acute conjunctivitis, with a serous or seropurulent discharge. A frothy fluid exuded from the mouth, and if the body had lain unmoved for several hours, a pool of coagulated fluid was seen about the head. Post-mortem changes were conspicuous even within

CHART XXIV.- Duration of life after chlorine gassing

3 to 4 hours after death, and decomposition could be quite advanced after 5 or 6 hours. An invasion of the tissues by the gas bacillus was often seen, and the liver, even after only 2 or 3 hours, was at times mottled with gaseous foci of varying size. That autolytic processes proceeded more rapidly after death in gassed than in nongassed animals seemed very evident, although no careful or systematic observations were made regarding the phenomenon, which was apparently an acceleration of the normal autolytic processes.

The abdominal organs showed a pronounced congestion, which was particularly striking in the liver. The congestion was less marked in the kidneys and spleen. The vena cava and its larger branches were distended. This splanchnic engorgement undoubtedly was dependent upon disturbances in the pulmonary circulation, discussed below.

On opening the thorax, the voluminousness of the lungs, which tended to overlap in the median line was very striking. The pleural surface did not show


the usual wrinkling of the normally collapsed organ, but was tense, glossy and semitranslucent. The tissues in the anterior mediastinum, including the areolar tissue about the thymus and the great vessels were markedly edematous.

The heart was quite regularly distended with blood; sometimes markedly so. Occasionally the apex was bifid and the tip of the right ventricle might project further than the left. All chambers of the heart shared in this apparent dilatation, but the right ventricle seemed most affected. If the animal was autopsied immediately after death, the blood was fluid, but clotting took place rapidly, with the result that the clots were generally of the dark cruor type. This was true, particularly, of animals that had survived gassing only a few hours. Where death was delayed more than 12 hours, the clots were more often of the usual chicken-fat quality, indicating less rapid coagulation. The endocardium, as a rule, was quite smooth and pale, but occasionally on both the valvular and mural endocardium brilliant flame-like hemorrhages were found which might be irregularly stellate or rounded in form. Very rarely the endocardium over these hemorrhages was roughened and, in a single instance, a large thrombus was attached to such an area in the region of the tricuspid valve.

As the trachea and bronchi were opened, a quantity of frothy fluid poured out. The mucous membrane of the trachea was reddened and the vessels were deeply injected. In animals surviving only a few hours, the surface was quite smooth, though somewhat opaque, with a loss of the normal gloss and translucency. Later the surface was covered by a sticky, membranous exudate which was quite tenacious, and was removed with difficulty. The gross appearance of the larger bronchi was practically identical with that of the trachea.

The lungs were very voluminous and retained their shape upon removal, even after considerable fluid had escaped through the bronchi. They collapsed slightly after sectioning, chiefly from the loss of fluid through the several bronchi. The tissue was more or less doughy throughout, with faint crepitation in places. The color was brilliant and variegated. The background was a red of deep, rich quality, with a mixture of blue, giving to the whole in general, a purple hue. There were lighter-colored patches, most numerous toward the margin of the upper lobes but scattered generally through all lobes. These paler areas were slightly elevated and were readily recognized as patches of emphysema, obviously compensatory, as in human lobar pneumonia. In some places the air of the distended vesicles had been replaced by fluid. Occasionally dark red depressed patches were seen, which clearly represented partially collapsed lobules.

When the lung was sectioned a tremendous amount of fluid escaped. The cut surface was red and more or less translucent. There was very little air in the bronchi and still less in the lung tissue proper. The escaping fluid was generally rich in albumin and might coagulate on standing. The larger blood vessels stood out prominently everywhere, owing to the presence about each of a relatively white edematous zone which might measure as much as 4 mm. in width. In animals dying within a few hours after gassing there was no gross evidence of an inflammatory reaction in the lung, other than the presence of the serous exudate just mentioned: but in animals that survived 12 to 24 hours small areas of pneumonic consolidation were sometimes recognizable


with the naked eye, although with difficulty, owing to the coexisting edema and congestion which obscured the picture. A well-developed pneumonia was not often seen in the first 24 hours.


In the respiratory tract, where practically the only changes of importance were found, histological examination not only confirmed at once the gross findings in the way of necrosis of tracheal epithelium, pulmonary edema and congestion, focal emphysema and atelectasis, but it also threw some light on the mechanism of the development of these lesions.

The changes in the epithelial lining of the respiratory tract, which were undoubtedly fundamental, will be considered first. (Fig.48.) The mucosa of the trachea and bronchi, even in cases of death in two hours after exposure looked completely coagulated. (Fig.49.) The nuclei were pycnotic and the cytoplasm homogeneous and glassy. The membrane was often raised in blisterlike fashion from the submucosa, and sheets of it were found lying in the trachea and bronchi. (Fig. 50.) It soon began to slough, and fragments free in the lumen underwent rapid disintegration. This epithelial necrosis was is clearly seen even in the smallest bronchi, but in the atria and air vesicles the damage to the inconspicuous lining cells was not so evident, although there was more or less desquamation.d

FIG. 48.- Normal bronchus. Bronchiolar termination in a dog's lung, showing transition from high ciliated epithelium of bronchi to the flattened epithelium of the infundibulim.

A thorough examination of the lung showed that the damage was not limited to the destruction of the lining epithelium, but that in many instances foci of hyaline necrosis of the entire alveolar wall were found. (Fig. 51.) These focal necroses varied in size, but generally comprised a group of air vesicles

d. In order to determine the extent and rapidity of the damage of the epithelium, resort was made to use thorough examination of the stains, which, as is well known, serve to distinguish dead from living cells, before catabolic changes have taken place, no such distinction being possible by ordinary staining methods. Dogs were given appropriate doses of trypan blue intravenously, and after a proper interval were subjected to a lethal exposure of chlorine. Some were killed shortly after gassing; others died within 24 hours. In all cases examination of the respiratory tract showed an intense nuclear stain, not only of the lining cells of the trachea and bronchi, but also of much of the flattened alveolar epithelium, particularly that about the atria. From these observations it was concluded that chlorine acted directly on the respiratory epithelium and that cell death immediately followed exposure. With this action of the gas clearly in mind, many of the other pathological phenomena, particularly the bacterial invasion and the associated massive inflammatory reaction, became readily explicable, as will be emphasized in the subsequent discussion.


which in some sections were seen to open into a common bronchiole. (Fig. 52.) The alveolar walls were homogeneous, glassy, and stained deep pink in hematoxylin-eosin preparations. In our experience this lesion was not associated with any other gas than chlorine. In a careful study of the lungs, in 50 acute deaths from chlorine, areas of necrosis were demonstrable in all but 3, and in these only two sections from each case were examined. It would appear then that the association of the lesion with chlorine was sufficiently constant to make its presence of considerable differential diagnostic value in warfare where the type of gas used is not always known.

The pulmonary capillaries, as the gross picture indicated, in most places were engorged with blood. This was particularly true in the partially collapsed areas. In general, the content of the vessels was not obviously altered. The leucocytes were sometimes slightly increased in places, but this was not marked. Thrombi in the pulmonary vessels have been described by several investigators including Klotz 1 and Bunting, but in the studies here recorded it was not satisfactorily determined that in dogs, at least, definite thrombi were present,

FIG. 49.- Bronchus plugged with sloughs of "cooked" epithelial lining. The cytoplasm of the cells is quite homogeneous and hyalinized

either in the capillaries or larger vessels, nor was it clear that the picture which Klotz described was sufficient evidence of antemortem coagulation. With special stains, however, it was possible to demonstrate, in many cases, a thick layer of fibrin covering the alveolar wall, and in places fibrin strands crossing the capillary to the alveolar surface of the adjacent air chamber. Presumably these fibrin deposits were laid down in the course of the outpouring of plasma from the capillary into the alveoli; in other words, the plasma coagulated partially or completely as it exuded. It was not only conceivable but very probable that such Ia coagulum would impede the flow of blood through the capillaries, even if it did not lead to intra vitam clotting, as Klotz1 maintained. The blood in the larger vessels of the lung looked normal and the intima and media of the vessels were unaltered. The adventitia, however, was in many cases strikingly edematous and the perivascular lymphatics, as Edkins and Tweedy 2


have emphasized, in their experiments with rabbits, are markedly distended. Occasionally, the edema of the perivascular sheath was replaced by blood, resulting in a circular, ring-like hemorrhage.

The outpouring of fluid not only into the air sacs of the lung but also into the tissue spaces of the tracheal, bronchial, and alveolar walls constituted both grossly and microscopically, as it did clinically, the dominant feature of the pathological picture. In some cases the pulmonary alveoli were filled with the precipitated pink-staining fluid, making the organ resemble superficially the thyroid with its colloid-filled vesicles. The fluid varied in its albumin content, as indicated by the slight precipitation in some cases. This variation might

FIG. 50.- Higher magnification of necrotic lining shown in Figure 49

be due to the stage of the edema, the fluid coming out first being, according to Klotz,1 poor in albumin, or it might depend on the degree of permeability of the damaged alveolar walls. The interstitial tissue, as well as the air sacs, was filled with fluid. The adventitia of the larger blood vessels was usually strikingly edematous, with a dilatation of its lymphatics. (Fig. 53.)

In addition to the edema there were seen quite early scattered mononuclear cells in the alveoli. These were for the most part desquamated epithelium, but there were a few mononuclear and polymorphonuclear leucocytes. In a certain number of cases there were accumulations of polynuclears, particularly in and about the atria, giving the picture of an early pneumonia. (Fig. 54.) It is a question as to whether this focal reaction was due to the gas injury or to invading


pathogenic bacteria. This question will be discussed in a subsequent paragraph, but it may be stated here that the investigations tended to show that once the protective epithelial barrier was destroyed, bacteria entered the lung almost immediately and that these focal reactions were undoubtedly the results of such invasion.


Reference to Chart XXIV shows that of the animals which came to autopsy 172 died during the first 24 hours, 25 on the second day, and 8 on the third day. These figures emphasize the fact that the bulk of the fatalities occurred during the first 24 hours, but it is obvious that any line of division between these acute deaths and what may be termed delayed deaths must be quite arbitrary.

FIG. 51.- Multiple areas of focal necrosis in lung of dog dying 20 hours after chlorine gassing

Certainly as regards pathologic findings, there was no abrupt transition. Furthermore, among animals dying in the first 24 hours, pneumonia, if present, was seen only as an early lesion, entirely obscured by edema and congestion, whereas in the later deaths there was a well-marked pneumonia in at least 95 percent of the cases There is here, therefore, a fair basis of division.

TABLE 50.- Dogs gassed with chlorine


As might be expected, the pathological findings in these delayed deaths showed a greater variation than was noted in the acute cases. The time factor, of course, was important. In animals dying 2 to 4 days after exposure, the pulmonary edema and congestion which reached its maximum during the first 24 hours was still present to a marked degree, though it might be on the decline. Pneumonia, as stated above, was practically always present. Grossly, it may be either of the lobar or lobular type, but if the former, the appearance suggested a confluent lobular rather than a true lobar process. The relation of the bronchi to the patches of consolidation was usually quite evident. Histologically, there

FIG. 52.- Higher magnification of an area of focal necrosis such as is shown in Figure 51. Death occurred six hours after gassing. The alveolar walls about an atrium are completely hyalinized and stain deeply with eosin

were special features of the pulmonary and bronchial lesions which had a certain resemblance to the changes found in the lung in epidemic influenza. There was, for example, the same striking degree of injury of the bronchial wall, and the focal hyaline necrosis of the alveoli. There was also, as in influenza, an active regeneration of the respiratory epithelium which was seen not only in the alveoli but even more strikingly in the bronchi. As previously described, the gas destroyed the ciliated epithelium of the trachea and bronchi. There was left, however, the deeper syncytial layer, the cells of which might be seen in an active state of multiplication as early as 48 hours after gassing, and the process became still more active after 4 or 5 days. Several mitotic figures might be found in a


single field. In the smaller bronchi, however, the distribution might be so complete that even the deeper laver of epidermis was destroyed, and repair took place only by a new growth of connective tissue which filled the lumen and eventually obliterated it. (Fig. 55.) A widespread organizing pneumonia was not infrequent in animals living 5 to 10 days. Abscesses and small gangrenous cavities were also common complications. In some instances the gangrene evidently originated in foci of necrosis such as were described in connection with acute deaths, but in most cases the lesion was undoubtedly referable to the action of bacteria. Suppurative pleurisy was an infrequent complication.


In any large series of animals receiving an average lethal exposure to chlorine a certain percentage of recoveries will be observed. As Underhill

FIG. 53.- Marked perivascular edema in acute death from chlorine. The lymphatic channels of the adventitia are widely distended with fluid

and others have pointed out, this was true of practically all of the toxic war gases, and was due partly to individual variations in susceptibility and partly to the technical difficulty of insuring a uniform dosage for each animal. If the recovered animals were kept under observation it would be found that some died sooner or later as a result of changes referable directly or indirectly to the gassing, and that many of the others, if killed, showed secondary effects of the gas injury. The percentage of cases in which perfectly normal organs were found at autopsy in such a group of animals was relatively small. The following summary of observations in a large number of "recovered dogs" will serve to confirm the general conclusions just stated.

There were 79 dogs living 15 days after gassing. Of these, 26 died between the fifteenth and one hundred and ninety-third days, when the observations ended. Of these deaths, 11 occurred between the fifteenth and thirtieth day. The remaining 53 animals were killed, most of them between 3 and


6 months. In nearly all cases, a few days after gassing, the dogs were sent to a farm in the country, where they were kept until they died or were killed. Many of them looked to be in good health, but some were lean and sluggish, and began to cough when made to exercise freely. Those that died were poorly nourished, without exception, and a few were quite emaciated. All showed at autopsy well-marked anemia, moderate fat accumulations in the liver and kidneys, enlargement of the spleen, in addition to the chronic changes in the lungs, which will be described in some detail.

The lungs showed in some cases maroon-colored areas of partial or complete atelectasis, while the rest of the pulmonary tissue was irregularly emphysematous

FIG. 54.- Early inflammatory reaction in the lungs found six hours after exposure to chlorine. The cells of the exudate are chiefly polynuclears which accumulated in and about an atrium

The atelectatic areas were associated quite regularly with gross changes in the bronchi leading to these parts. These changes were chiefly a thickening of the walls and plugging of the lumina, with a mucopurulent exudate. The extent of these changes varied in individual cases. Microscopically, the emphysema and atelectasis noted grossly were quite evident.The associated changes in the bronchi varied. In places there was a pronounced chronic infection, the lumen being filled with leucocytes and cellular debris. The mucosa and other coats were infiltrated by wandering cells and in some instances a wide mantle of cells in the adjacent lung tissue. (Fig. 56.)


Another and more characteristic lesion was found in the smaller branches of the bronchial tree. This was what might be termed an organizing or obliterative bronchiolitis. In the development of this lesion there was evidently first a partial and later a complete occlusion of the bronchiole. The lumen was first filled with an exudate composed of leucocytes, fibrin, and necrotic epithelium, as seen in some of the delayed deaths described above. This exudate became gradually organized through the ingrowth of blood vessels and fibroblasts from the bronchial wall. (Fig. 57.) A noncellular fibroid scar might finally result, or in some instances the organized mass might persist as a sort of polyp partly covered by epithelium. (Fig. 58.)

FIG. 55.- Organizing bronchiolitis five days after chlorine gassing. Lumen of bronchus is filled with a network of fibroblasts and there is a similar organizing process going on in the adjacent pulmonary alveoli. In the bronchial and alveolar walls there is much old hyatinized fibrin

This partial or complete bronchial obliteration explained the presence of the patchy emphysema and atelectasis to which reference has already been made. It may be mentioned in passing that these chronic bronchiolar lesions were quite analogous to those observed in the human lung, as described by Wagner3 an others.

A considerable number of animals showed, in addition to the healed bronchial lesions, with resulting mechanical disturbances, evidence of a superimposed chronic infection. In these cases chronic bronchitis, bronchiectasis, and patches of organizing pneumonia were observed, the extent of the change


being dependent apparently on the type of infecting organism and the amount of initial gas damage, though of course the undetermined factor of individual resistance had to be taken into account. The character of the commoner bronchial lesions is shown very well in Figures 58 and 59. In a few instances, the infection had spread to the pleura with the development of a typical empyema, and there were several cases in which a generalized infection had taken place, with metastatic abscesses in the viscera. These instances of extensive infection, however, were exceptional, the usual picture being simply a chronic purulent bronchitis with more or less peribronchial reaction.

Routine gross and microscopic examination of various organs and tissues, including intestinal tract and brain, disclosed no chronic change of any significance. It seems reasonable, therefore, to assume that the poor general condition of the animals was referable to impaired pulmonary function, although it must be granted a much more exhaustive study of this question is necessary

FIG. 56- Purulent bronchitis in a dog dying 23 days after exposure to chlorine. The cells in the bronchi are chiefly polynuclears, but in the adjacent alveoli are many large mononuclear cells before any definite conclusion can be reached. The problem is clearly an important one in that its solution would throw light on the nature of the chronic disability seen in many of the recovered gassed soldiers.


The inhalation of chlorine resulted in damage to the lining epithelium of the entire respiratory tract, the degree of injury depending upon the concentration of the gas and the duration of exposure. In lethal exposures, there was complete necrosis of the tracheal and bronchial mucosa with subsequent sloughing and focal necrosis of the pulmonary septa. When death occurred immediately, the injury was demonstrable by the use of vital stains; no inflammatory reaction was manifest. In deaths occurring a few hours after gassing, the initial injury was obscured by reactive phenomena, chiefly edema and congestion of trachea, bronchi, lungs. and interstitial tissues. Where death was further delayed, bacterial invasion of the damaged tissues occurred through


the loss of the normal protective mechanism of the upper air passages. This resulted in a widespread inflammatory reaction involving trachea, bronchi, and lungs.

There was no anatomical evidence of injury to tissues other than those of the respiratory tract. Dilatation of the heart and circulatory disturbances, when present, were undoubtedly secondary to the pulmonary lesions.

Animals recovering from severe exposures showed chronic changes in the lungs of the nature of organizing bronchiolitis, chronic bronchitis, bronchiectasis and emphysema. These changes were referable partly to the initial gas damage

FIG. 57.- Obliterating bronchiolitis in a dlog killed 32 days after chlorine gassing. The main bronchus has a normal, probably regenerated, epithelial lining. The small tributary bronchus is occluded by an organized mass of tissue adherent to the wall in places

and partly to the superimposed bacterial infection. They appeared sufficient to account for the general debility shown by some animals.


C-270.- Male mongrel; weight 24 kilos. January 3, 1918, gassed 30 minutes with chlorine, 858 parts per million. Died three hours after exposure.
Autopsy.- Tracheal and bronchial mucosa opaque, necrotic looking, and easily peeled off. Lungs extremely edematous and congested.
Microscopic examination shows a severe diffuse injury to the lining epit helium of the air passages, large and small, and focal necrosis of the lungs, with practically no cellular inflammatory reaction. Sheets of hyaline necrotic epithelium are, in most places, loosely adherent as in Figure 49, but some bronchioles are plugged with fragments of detached membrane,


as shown in Figure 60. The general engorgement of the pulmonary capillaries is the outstanding feature of the lung changes. Relatively little precipitated edema fluid is seen in the alveoli. The sheaths of the large blood vessels, are, however, extremely edematous with prominent dilated lymph vessels. Numerous foci of necrosis are seen involving the atria and the adjacent alveolar septa. Figure 50 shows very well the size and relations of such an area. Special stains show a thick layer of fibrin along the dead alveolar walls.
Anatomical diagnosis.- Necrosis of tracheal and bronchial mucosa; focal necrosis of lungs; extreme congestion and moderate edema of lungs.

NOTE.- The short duration of life after gassing (less than three hours) accounts for the absence of appreciable cellular inflammatory reaction. The case illustrates well the early changes following chlorine injury.

FIG. 58.- Chronic bronchitis and patchy emphysema in a dog dying 55 days after chlorine gassing

C-136. -Young male cocker spaniel. November 17, 1917, gassed 30 minutes with chlorine, 847 parts per million. Salivation and defecation during exposure, and evidence of marked respirators- irritation. Died eight hours after gassing.
Autopsy.- Apart from congestion of the abdominal organs, pathological changes are limited to the respiratory tract. The trachea and large bronchi show an opaque, grayish membranous lining in place of the normal velvety mucosa. The lungs are diffusely edematous, and deeply congested. There is also marked edema of the mediastinal tissues.
Microscopic examination shows complete necrosis and sloughing of the mucosa of the trachea and bronchi, with a widespread polynuclear reaction, and much edema of the entire wall. (Fig. 61.) The more superficial mucous glands are necrotic with an invasion of leucocytes. The injury and reaction in the smaller bronchi are equally well seen. The pulmonary capillaries are widely dilated throughout. There is very little albuminous precipitate or fibrin in the alveoli, and few free cells except in the atria and bronchioles. Here and there are typical foci of hyaline necrosis as shown in Figures 49 and 50.
Anatomical diagnosis.- Acute necrotizing tracheitis and bronchitis; edema, congestion and focal necrosis of lungs.

NOTE.-The findings are typical of acute chlorine poisoning. The inflammatory reaction was well advanced in the trachea and bronchi, although death occurred only eight hours after gassing.


C-134.- Male collie; weight, 16 kilos. November 16, 1917, gassed 30 minutes with chlorine, 867 parts per million. Died 24 hours after gassing.
Autopsy.- Lungs are voluminous, heavy, and deeply congested. In the trachea and bronchi there is a membranous exudate throughout and the walls of the air passages are quite edematous. No pneumonic patches are seen or felt.
Microscopic examination.- The most striking changes are found in the trachea and bronchi. The mucosa is necrotic and in most places sloughed off, leaving a denuded surface or only the basal layer of epithelium. (See Fig. 51.) The tracheal wall is infiltrated by leucocytes, with an outpouring of serum. The adventitia of the blood vessels is extremely edematous (see fig. 51), with widely dilated lymph channels, filled with precipitated lymph. In the alveoli there are, in addition to the edema, a few free cells. Some of these are polynuclears, others are desquamated epithelium. There are no frank pneumonic foci. Several small areas of focal necrosis of the septa are seen, with more or less hemorrhage.
Anatomical diagnosis.- Necrotizing tracheitis and bronchitis; focal necrosis and hemorrhages in lungs; pulmonary and mediastinal edema and congestion.

FIG. 59- Bronchiectatic cavities in lungs of dog dying 39 days after chlorine gassing. The lung tissue supplied by these bronchi is completely atelectatic

NOTE.-In view of the widespread loss of the protective lining of the upper respiratory tract, with the evidence of tracheal infection, it is rather surprising that a bronchopneumonia had not developed. This would undoubtedly have occurred had the animal lived much longer. The early regenerative changes in the bronchial epithelium show the rapidity with which repair takes place.

C-109.- Brindle Male mongrei; weight, 18.2 kilos. October 31,1917, gassed with chlorine 30 minutes. Condition following gassing fair, but respiration became somewhat labored. Refused food and appeared depressed, but able to walk about. Condition remained about the same until death on November 5, five days after gassing.
Autopsy.- Body weight, 16.4 kilos, a loss of 1.8 kilos after gassing. The lungs are deeply congested, heavy, firm, and contain very little air.
Microscopic examination.- The lungs present an unusually complex microscopic picture, showing a confused mixture of inflammatory and reparative changes. In a majority of the


bronchi the ciliated surface epithelium is lost, but in many places the remaining basal cells have proliferated to form a thick layer the thickness of three or more cells. Mitotic figures are easily demonstrable. (Fig. 62.) Along with this reparative process, there is evidence of acute infection, with masses of polynuclear leucocytes in both the bronchial lumen and wall. It is noteworthy that occasional bronchi of moderate size are found showing a practically normal ciliated mucosa. The lung parenchyma shows much the same changes as the bronchi-active regeneration of alveolar epithelium, interstitial fibroblastic proliferation, particularly about the bronchioles, fibrinous and cellular alveolar exudate, miliary abscesses, and hemorrhages. Specific stain shows a thick layer of fibrin on the alveolar wall in many places. There is active proliferation of the surface cells of the pleura, with a prominent deeply staining layer.
Anatomical diagnosis.- Necrosis and regeneration of respiratory epithelium (tracheal bronchil, alveolar); suppurative and organizing bronchopneumonia.

FIG. 60.- Sloughing of bronchial epithelium in dog killed three hours after exposure to chlorine

NOTE.-The case illustrates well some of the sequelae of a moderately severe exposure to chlorine. The marked variation in the injury suffered by different bronchi is difficult to explain except on the assumption that the less injured bronchi had been spasmodically contracted during exposure.

C-334.- Male setter; weight 10 kilos. January 17, 1918, gassed 30 minutes with chlorine, 716 parts per million. Recovered and sent to farm January 31, where he remained until May 14, approximately five months after gassing, when he was killed. Condition at this time was recorded as "somewhat emaciated and mangy.''
Autopsy.- Pathological changes are limited to the longs which are described as follows: The lower lobes are grayish pink and fairly well collapsed. The anterior margins of upper lobes are dark purple and airless, firm and nodular. Elsewhere in the upper lobes are light-colored emphysematous patches. On section the bronchi in both upper and lower lobes are conspicuous by reason of their thickened walls. In the upper lobes they are much dilated and filled with mucopurulent exudate. In the atelectatic areas along the anterior margins there is obviously a marked connective tissue overgrowth.


FIG. 61.- Acute necrotizing bronchitis 10 hours after chlorine gassing. There is complete sloughing of mucosa and a diffuse inflammatory reaction throughout the bronchial wall. Note the marked edema of tissue surrounding a large peribronchial vessel


Microscopic examination.- A number of sections from different portions of the lungs show very much the same pathological changes, namely, thickening and mononuclear infiltration of the bronchial walls, bronchiectasis, purulent bronchitis, patches of organizing pneumonia, focal obliterative bronchiolitis. (Figs. 63, 64, and 65.) Gram-positive diplococci are found in the bronchial exudate.

NOTE.- The widespread and advanced inflammatory changes in the bronchi and lungs are quite adequate to account for the poor condition of the animal when killed. Since the lesions were apparently progressing, the dog would no doubt have succumbed within a comparatively short time.

C-55.- Young male collie; weight, 14.4 kilos. September 28, 1917, gassed 30 minutes with chlorine, 678 parts per million. Stood gassing well and after a few days appeared to be in normal condition. Regassed on October 20, 22 days after first exposure, 30 minutes; concentration 710. Stood gassing poorly, and though the acute period was passed without

FIG;. 62.- Regeneration of bronchial epitheliuns four days after chlorine gassing. One mitotic figure is seen

marked respiratory disturbance, a cough persisted. Killed with chloroform November 6, 39 days after first gassing and 17 days after second gassing.
Autopsy.- Body poorly nourished and anemic. Lungs are irregularly collapsed. The most outspoken changes are in the bronchi, many of which are greatly dilated and filled with pus. In other parts of the lung there are foci of obliterative bronchiolitis, with radiating scars, and more or less emphysema.
Anatomical diagnosis.- Chronic bronchitis and bronchiectasis; obliterative bronchiolitis; focal atelectasis and emphysema; anemia.

NOTE.- The case shows unusually advanced lesions involving primarily large and small bronchi. The changes were no doubt intensified by the second gassing, and would probably have ultimately caused the death of the animal from a flare-up of the persistent respiratory infection.


The mode of action of phosgene accounts for the localization and character of the chief anatomical changes resulting from its inhalation. The toxicity of the gas, as has been pointed out elsewhere, is due to the fact that in the presence


of water it is split up into hydrochloric acid and carbon dioxide. When inhaled, little decomposition takes place until the gas reaches the lungs where, in the small air passages and sacs, it comes into contact with sufficient water vapor to bring about the evolution of hydrochloric acid. According to Hoover, 4 much of the phosgene taken into the lungs is probably absorbed as such, decomposition taking place gradually in the tissues. This would explain the severe damage to the bronchiolar and alveolar epithelium, and would account for the more gradual development of the signs of pulmonary injury than is the case with certain other gases of the respiratory irritant group, such as chlorine and bromcyanogen, which do not require for their action any preliminary decomposition. It may be observed, in this connection, that although the phosgene injury is referable to the hydrochloric acid formed from it, the effects of phosgene and hydrochloric acid inhalations are by no means identical. On the contrary, the lesions produced by the two gases differ both in extent and degree. Hydrochloric acid injures most severely the larynx and trachea, damaging less the distal portion of the respiratory tract, while in the case of phosgene poisoning, it is the distal portion which is most affected. The

FIG. 63.- Law magnification of Iung of dog dying two months after exposure to chlorine. Chronic bronchitis and bronchiectasis with atelectasis and occasional patches of emphysema

importance of the locus of decomposition is further emphasized by the observation that if animals are exposed to hydrochloric acid and phosgene in equivalent concentrations, the former is found to be distinctly less toxic, probably because its action is expended on the less vital proximal portion of the respiratory apparatus.

Extensive and detailed studies of the pathological changes induced by phosgene on experimental animals have been made both in this country and in England, but since the chief purpose of this chapter is to review the work done in America, the studies carried on by the medical section of the Chemical Warfare Service will be presented in some detail, with a brief review of the findings of British workers.

Investigations were conducted at both American University, Washington, D. C., and Yale University, New Haven, Conn. At the former station, a comparative study of the lesions in various animals was made, while at New Haven a systematic study of the changes in dogs, from the earliest to the most chronic lesions, was made, using animals gassed in connection with the studies upon the mode of action of phosgene and the treatment of the condition produced by toxic inhalations of the gas. 5


The dogs were exposed in closed chambers for 30 minutes to concentrations varying from 44 to 120 parts per million. Although with higher concentrations there was, as has been pointed out in Chapter X, a more rapid onset of symptoms and a larger percentage of fatalities, the changes in the respiratory tract, as far as it has been possible to determine, were practically the same with high and low concentrations, except in respect to the time required for the development of the lesions.


The animals dying after gassing were autopsied, in most cases within a few hours. Dogs which did not succumb within two or three days (designated "recovered animals") were sent to a farm where a certain number died sooner or later (Table 51). The remainder of the recovered dogs were killed at intervals up to 129 days after gassing, and complete autopsies were performed.

FIG. 64- Higher magnification of the bronchi shown in Figure 63. Bronchus is moderately dilated. Adjacent lung tissue is atelectatic

In order to study the very earliest changes, a few animals were killed shortly after gassing, before the development of marked symptoms.

As might be expected, the changes found at autopsy, after fatal gassing, varied with the length of time the animal had survived. For this reason it will make for clearness of description if the findings in the several stages, which may be conveniently termed "acute," "subacute," and "chronic," are discussed separately.

These periods are arbitrarily divided as follows: (1) Acute, death within 48 hours; (2) subacute, death 3 to 10 days; (3) chronic, survival of 10 days or more. The accompanying table shows the number of animals falling into each of these groups and, in addition, the deaths by days up to the fifteenth day. Omitting the `"killed animals," it is seen that in the majority of cases death occurred within the first 24 hours, and that the number of deaths in


each successive 24-hour period diminished progressively until the tenth day, when the base line was nearly reached (Chart XXV). It should be clearly stated, however, that the figures given in the table, and the curve based on them, refer only to autopsied animals, and do not include all of the dogs from any single experiment or series of experiments. In other words, the figures are intended to show only in a rough way the time at which death is likely to occur after phosgene gassing.


Out of 503 animals gassed 376 died, of which 256, or 68 percent died within the first 48 hours. Analyzing the figures further, we find that most of the deaths (195 of the 256) took place between 12 and 24 hours after gassing.

FIG. 65.- Higher magnification of two bronchi shown in Figure 63. Lumina are filled with a cellular inflammatory exudate. There is also a chronic peribronchial reaction

there being only 29 in the first 12 hours, and 32 between 24 and 48 hours. In other words, the critical period, as Underhill and others have shown, lay in the second 12 hours after gassing.

TABLE 51.- Dogs gassed with phosgene


TABLE 52.- Dogs gassed with phosgene

The gross anatomical changes found in animals dying acutely, though very striking, showed little variation. After the first, second, or third autopsy, the impression was one of uninteresting sameness, relieved only by the remarkably severe and brilliant changes in the respiratory system.

CHART XXV.- Duration of life after phosgene gassing

Frothy fluid, generally clear, but occasionally blood-tinged, oozed from the mouth. The conjunctivie might be slightly reddened, but a frank conjunctivitis was rarely seen. Decomposition proceeded rapidly after death, as in chlorine poisoning, and was quite marked if the autopsy was delayed more than three or four hours, unless the body was kept chilled. The body weight was regularly less than at the time of gassing. This difference was often considerable, as the appended protocols show, amounting in the case of a large dog to as much as 1.5 kilos. The loss was probably due, in large part at least, to the escape of fluid from the edematous lungs through the mouth.

The engorgement of the great vessels and the congestion of the abdominal viscera were quite pronounced. The liver was swollen and extended well below the costal margin. Its dark purplish color was partly lost as the vessels were cut, allowing the blood to drain out. The spleen was only slightly enlarged but, like the liver and other abdominal organs, it presented the picture of acute congestion.


As the sternum was removed the voluminousness and noncollapsibility of the lungs were at once evident. The anterior margins nearly met in the median line and tended to overlay the heart. The pleural cavities were obliterated and it was unusual to find more than a few cubic centimeters of fluid in either the pleural or pericardial cavities. The loose tissue of the anterior mediastinum, however, as well as the interlobar pulmonary septa, were generally more or less edematous. The heart, as a rule, was quite distended. The distention was generally more pronounced on the right side. (Pl. XI.) In one animal that succumbed seven hours after exposure and was autopsied immediately, the apex of the heart was bifid by reason of the marked right-sided dilatations.e The cavities and valves of the heart presented nothing abnormal except for the presence of flamelike hemorrhages beneath the endocardium, which were encountered not infrequently. They were found in approximately 15 percent of the acute deaths. The commonest location of the hemorrhages was in the region of the papillary muscles of the left ventricle.


Respiratory tract.-The larynx might be slightly edematous, but was otherwise normal looking. The trachea and bronchi were filled, more or less completely, with frothy fluid. The fluid was generally clear, faint yellow, and of the consistence of blood serum. In the small bronchi it might be slightly viscid, or even tenacious, owing to an admixture of mucus and fibrin. Hill 6 has pointed out that if the fluid is collected it may often coagulate on standing but this phenomenon was not observed. The mucosa of the trachea and larger bronchi was usually of normal smoothness and color, except that between the cartilaginous rings it was at times slightly reddened from injection of its vessels, which presented, on close inspection, a fine network. The congestion was more marked in the smaller than in the larger bronchi.

The lungs were strikingly voluminous and quite heavy. The pleura was smooth, glistening, and moist, and its lymphatics were often distinctly dilated. In this early period a fibrinous or purulent exudate on the surface was practically never seen. The lungs were extraordinarily mottled; large whitish patches alternating with deep red ones. Over the whole, there was often a bluish cyanotic hue. The light-colored bodies were more or less elevated and crepitant, and were easily recognized as areas of acute emphysema. By reason of the distension, the individual alveoli could often be made out with the naked eye, and could be seen with great clearness by aid of a low-power magnifying glass. The darker portions of the lung appeared collapsed by way of contrast, but examination showed that these areas were not particularly atelectatic, but only less distended than the remainder of the lung tissue. The dark color was referable to the extreme congestion. The proportion of light-colored emphysematous lung tissue and dark congested, partially collapsed lung varied considerably. In some cases almost the entire lung was dark bluish red, with only a few scattered light patches in the upper lobes. In other cases, where the duration of life after gassing was the same, a widespread emphysema, with little of the (lark red tissue, was found. In general the lower lobes were darker, the emphysema being more pronounced in the upper and middle portions.

e The phenomenon of cardiac dilatation in gassed animals is discussed in some detail in subsequent pages.


On palpation the lung was doughy and pitted on pressure. Much fluid could be expressed through the bronchi. In animals dying in the first twelve hours, the edema was generally considerably less than in those living longer. In the latter, it was generally the most conspicuous feature, quite overshadowing the congestion and emphysema. In extreme cases, the rounded margins of the lung were quite translucent and the dark bluish-red color of the early period was transformed into a watery pink, as though even the stagnant blood in the vessels had become diluted.

In estimating the extent of the edema, the weight of the lungs, as has been pointed out in connection with chlorine gassing might be taken as a fair index, although it was evident that the greatly increased weight of the organ was due not to the edema alone, but in considerable part to the extreme congestion which was regularly present. Indeed, in the first few hours before the edema became well developed, the excess of blood was clearly responsible for the larger share of the increase in weight. Both the heart-lung and body-lung ratios were used as indices of edema. The former was suggested by Barcroft, and has been used by English observers. The normal indices were determined from autopsies on 15 normal dogs, and an average lung-heart ratio of 1.3 and lung-body-ratio of 0.0115 being found. Dividing the index obtained in the gassed animals by these normal ratios, a figure was obtained which probably represented fairly accurately the degree of edema present. Of the two indices, that based on the lung-body ratio was regarded as the more accurate in the average, medium-weight, well-nourished dog, but in small animals it often gave readings which were obviously too high. An objection to the lung-heart proportion is that even slight variations in the method of trimming the heart will lead to relatively large errors. Using the two methods, it was found that the lung increased rapidly in weight after gassing, reaching a maximum after about 24 hours, when it might be, in fatal cases, more than four times the normal. (For detailed figures see Table 53, Chart XXVI).

TABLE 53.- Degree of increase in lung weight after gassing

While the cut surface of the lung was generally smooth and moist, close inspection might show tiny gray foci of bronchopneumonia. These were often overlooked in the gross specimen, being masked by the edema and congestion. The smaller bronchi were somewhat more conspicuous than normal owing to the edema of their walls.

Grossly, the blood vessels were quite normal looking. No thrombi were found in veins or arteries. Bronchial lymph nodes were somewhat enlarged and on section were distinctly edematous.


Light-colored patches of emphysema alternate with deep red congested and partially collapsed areas in the voluminous lung. Heart is dilated, particularly the right side.


Outside the respiratory system no pathological changes of any significance were found, other than the congestion of the abdominal viscera, and apparent dilatation of the heart already mentioned. In the stomach and intestines congested areas and hemorrhagic erosions were occasionally seen, but since similar lesions were not infrequently encountered in nongassed dogs, they were considered in no way related to gassing.

In 15 dogs the brain was examined, but in none could there be demonstrated capillary hemorrhages or inflammatory lesions such as were described by Mott 7 in human cases.

Examination of the pancreas, adrenals, and thyroid was in all cases negative.

CHART XXVI.- Degree of pulmonary edema after phosgene gassing, as determined by lung-heart and lung-body ratios


Significant changes were found only in the lungs and smaller bronchi. The alveoli varied markedly in size, some being exceedingly large, with very thin, bloodless walls, while others showed partially collapsed walls containing greatly distended, tortuous capillaries. There might be considerable desquamation of the alveolar epithelium, particularly in the semicollapsed patches.

The alveoli nearly everywhere contained more or less coagulated serum (edema), which might appear either as a homogeneous pink-staining material resembling thyroid colloid, or as a faint granular precipitate, the appearance depending apparently not so much on the quantity of fluid present as on its content.

In addition to serum there was generally some fibrin in the alveoli; occasionally this was quite abundant. But it was along the alveolar walls that


fibrin appeared in greatest amount. With special stains, a thick layer lining the alveolus could be demonstrated in many places, with frequent strands crossing the capillary to a similar layer on the wall of the adjacent alveolus. (Figs. 66 and 67.) It was evident that the presence of fibrin in this situation must not only interfere with gaseous exchange but, as will be emphasized in the discussion of the circulatory disturbances, also must constitute a serious obstacle to the flow of blood through the lungs. As in chlorine gassing, variations in the lumina of the cartilage-free bronchi were seen, which suggested alternate dilatations and constrictions, but one could not be certain that such pictures represented a condition that was not referable to post-mortem changes, such as irregular collapse of the lung. This point is discussed elsewhere (p. 457).

FIG. 66.- Lung of dog dying two days after exposure to phosgene. Fibrin stains show a heavy deposit along the alveolar walls, outlining them everywhere quite distinctly

The very large bronchi, like the trachea, showed little change, though there might be an excess of mucus and some polynuclear leucocytes in their lumina.

The epithelium was perfectly preserved, even the cilia being seen as distinctly as in the normal animal. The finer bronchial tubes, however, showed evidence of serious damage.

As early as two hours after exposure there was histological evidence of necrosis of the epithelium. The surface was covered by a thick layer of eosin-staining, material made up of mucus and dead desquamated epithelium. Beneath this necrotic layer there might remain a thin layer of flattened or rounded basal cells. In some cases these basal cells were destroyed, the tube being lined by a pink-staining, necrotic membrane.


The difference in the injury suffered by the proximal and distal portions of the respiratory tree was well shown by the use of vital stains. Ten dogs were injected intravenous on two successive days with 100 c.c. of 1 percent solution of the dye, and then exposed for a half hour to a concentration of phosgene varying between 80 and 97 parts per million. One animal was killed after two hours, and others at varying intervals up to three davs. Frozen sections, counterstained with carmine, were made from different portions of the respiratory tract.

In no instance was the tracheal epithelium vitally stained. The bronchial epithelium, on the other hand, was stained in places, and in the finer bronchioles the coloration was marked, affecting the entire wall quite uniformly. All

FIG. 67- Higher magnification of an area shown in Figure 66. In places the fibrin stain extends across the septa

of the bronchioles, however, were not equally affected. Some showed unstained epithelium, whereas the lining cells of other bronchioles in the immediate vicinity were deeply colored. The staining seemed most marked where there was distortion of the tube with either contraction or dilatation. The flat alveolar epithelium seemed unaffected. These results were obtained in animals that were sacrificed as early as two hours after exposure.

In addition to the necrotizing effects just described there were often seen, even in early deaths, a beginning inflammatory reaction. The reaction was practically always focal at this stage, with its point of origin in a bronchiole. (Fig. 68.) There was evidence, however, of what may be considered a general inflammatory reaction in the lungs. Nearly everywhere there was an increase in the number of polynuelear leucocytes in the alveolar walls and occasionally


they were seen in process of migration into the alveoli and fibrin; sometimes a considerable amount might be found in the alveolar exudate. The extent of the focal pneumonic process varied in different animals. Many showed no reaction at all.

A tabulation of the cases studied shows (Table 54) that pneumonia was demonstrated in approximately 50 percent of the animals dying in the first 48 hours.

TABLE 54.- Pulmonary complications in dogs gassed with phosgene


The content of the blood vessels in the lung appeared to be, in all cases, simple post-mortem clot, which, however, was strikingly rich in fibrin. (See fig. 67.) The question of the presence of fibrin thrombi in the capillaries described by Klotz 1 and others is discussed elsewhere. The extension of fibrin masses through the alveolar walls and capillaries an(l its importance has been referred to already.

The adventitia of the larger vessels was spread apart by edematous fluid, as in chlorine poisoning. (See fig. 53.)

The bronchial lymph glands showed a dilatation of their peripheral sinusoids, which contained mononuclear cells, red blood cells and occasional polynuclear leueocytes. The channels through the glands were often spread apart much more widely than usual and showed a similar cell content.

In the liver the hepatic veins were everywhere distended, and the engorgement was noticeable also in the capillaries which directly joined these vessels. In this central zone the liver cells were thinned out as though compressed, giving the typical picture of passive congestion as seen in man. There was generally no necrosis of the liver cells and no inflammatory reaction. In a few cases focal areas of necrosis were found, but on account of the rarity of this lesion we have considered it accidental and not related to the gassing. In the same way- we have interpreted a single instance of hemorrhagic cystitis and renal epithelial necrosis.


The important acute changes brought about by exposure to lethal concentration of phosgene gas were confined to the cardiorespiratory system. The upper respiratory tract was unaffected, and this was in marked contrast to the changes in the lungs and finer bronchi.

The lungs were the seat of an intense edema and congestion which was associated in many cases with focal inflammatory changes originating in the bronchioles. The inflammatory exudate was not confined to the bronchioles, but spread to a variable extent into the surrounding alveoli, so that a picture of early bronchopneumonia was found superimposed upon intense edema of the lung. Plugging of the bronchioles with exudate was associated with areas of partial atelectasis and emphysema of the lung tissue. The presence of an abundance of fibrin on and in the alveolar walls, crossing and obstructing the capillaries everywhere, offeredt an explanation for the increased resistance in the pulmonary circulation and the consequent dilatation of the right side of the heart.


There were, altogether, 66 dogs in this class, 15 of which were killed, and 51 died 3 to 10 days after gassing. The pathological picture, both gross and microscopic, was more varied at this period than in the earlier and more acute stage, where. as we pointed out, there was a certain uniformity, especially in the gross characteristics. On account of this variation, it will be necessary, in describing the findings, to subdivide the cases further into at least two classes: (1) Those that died; (2) those that were killed.

In the animals that died the anatomical changes were characterized by the presence of a severe and widespread inflammatory process in the respiratory tract. Thus Table 54 shows that in 90 percent pneumonia was found at autopsy and bronchitis in 96 percent.



The general appearance of the body, abdominal organs. and heart was practically the same as in the acute deaths. The chief differences were found in the lungs, which, in addition to edema, congestion, and patchy emphysema, showed quite regularly a more or less widespread inflammatory process, affecting both the smaller air passages and air sacs.

Grossly, the lungs were very voluminous and heavy. The surface was generally smooth and uniform, but very frequently firm, pale pinkish areas, often irregularly wedge-shaped, stood out in sharp contrast to the remaining portion of the lung, which was cushiony and crepitant. In animals that survived only three days, these solid areas were not as conspicuous as in the animals that lived longer. In practically all cases the consolidation was more extensive in the thinner lappets and near the margins of the lung. The posterior and dependent portions of the lung might be involved in the pneumonic process but, as a rule, parts which in the early deaths showed the greatest edema were not so regularly consolidated. It may be recalled in this connection, as previously noted, that the upper lobes are the more dependent in dogs and other four-footed animals.

Perivascular edema, as in the acute stage, made the larger vessels of the hilus conspicuous. In the vessels only post-mortem clots were found; no thrombi were present.

On section the lungs varied in their appearance according to the length of survival of the. animal. In the more acute deaths the lungs were still very wet, and it was sometimes difficult to make out the areas of consolidation which later stood out as dry granular areas that varied in size from a few millimeters to many centimeters. The dry granular areas were dark and reddish-brown in color, not nearly as translucent as the areas of edema and congestion, and very much firmer. In the smaller areas of consolidation there was almost always a central bronchiole on account of its thickened wall and the purulent exudate in its lumen. In animals surviving longer the areas of consolidation stood out much more strikingly and might involve the greater portion of a lobe, although as a rule only one-third to one-half of the lobe was affected. Not infrequently three or more lobes would contain extensively hepatized areas. The consolidation was generally pseudo-lobar in type but small discrete patches of bronchopneumonia might be seen scattered through the less affected lobes.


Histological studies corroborated the gross findings. With the exception of a few subsidiary lesions in the other organs, the changes were found only in the respiratory tract.

In all instances, the bronchioles were much altered. They were more or less dilated and appeared as large, round holes, filled with exudate, consisting of cellular debris, leucocytes, and red blood cells. In many places the lining epithelium was entirely lost and the walls were structureless. About such bronchioles there was generally an active pneumonic process, sometimes quite extensive and merging into other similar foci. In such areas, the alveolar walls might be necrotic, with the development of frank abscesses. The inflammatory exudate was often quite hemorrhagic. (Fig. 69.)


The pneumonic zone faded rather abruptly into areas where the lung tissue was well preserved. The alveoli here, as in the more acute stage, might be partially collapse(l or emphysematous. Their walls were always prominent on account of the dilatation of the vessels and the partial desquamation of the alveolar epithelium. The alveoli in this zone contained some fibrin and serum with occasional desquamated cells and leucocytes. That this was a very early inflammatory process was evident from the large number of polynuclear leucocytes caught in process of migration from the vessels.

The pneumonia was often widespread, approaching a lobar distribution, but it was fairly clear that the infection in these cases, as in the more patchy pneumonias, was bronchial in origin. Organization of the exudate was seen now and then, but was a much more common finding in the "recovered" dogs that were killed, as will be described later.

FIG. 69.- Bronchopneumonia causing death seven days after exposure to phosgene. The bronchial wall is necrotic and there is considerable hemorrhage in the pneumonic exudate. Lung is moderately edematous.

The bacteriology of these pneumionias was investigated in 23 cases, with the following results: Streptococcus hemolyticus was found alone in 10 cases; Staphylococcus aureus alone in 6 cases; Streptococcus and Staphylococcus together in 3 cases; Streptococcus and other organisms once in 2 cases; miscellaneous or undetermined organisms in 2 cases.

It is seen that. a hemolytic streptococcus was the most common organism met with, being demonstrated in 13 out of 23 cases. It was not possible to show a relationship between the type of organism present and the character of the pneumonia, although the, impression was obtained that abscess formation and organization were more often associated with the staphylococcus than with other organisms. It may be of interest to note that in a few cases of pneumonia among nongassed dogs, autopsied about the same time, the bacteriological findings, were roughly the same as in the gassed animals.


Microscopic findings, outside the respiratory system, were of little interest. In the liver the changes were very similar to those described in acute deaths. The congestion in the hepatic vein persisted and frequently the liver cells in the immediate vicinity were reduced to fine strands, with the intervening sinusoids greatly congested. These liver cells contained brown pigment, but very rarely was there any nuclear disintegration or leucocytic infiltration. The kidneys showed cloudy swelling of the tubular epithelium and occasionally the glomeruli would be markedly congested, but there was no evidence of any serious or permanent damage to the renal parenchyma.

FIG. 70.- Early stage of organization of pulmonary exudate four days after phosgene gassing


Fifteen dogs that were apparently recovering from the gassing were killed from the third to the tenth day for a study of the reparative processes in the damaged lung. Although the majority of the animals showed no symptoms at the time they were killed, the respiratory tract in all cases presented the obvious effects of the gassing. In most cases congestion, edema, and emphysema were still present in a moderate degree, being quite marked in some, though rarely so extreme as in the early fatal cases.

The most striking gross feature of the lungs was the presence of small nodules which were scattered quite uniformly throughout the organ. To the


palpating finger they felt very much like tubercles, though not quite so firm and shotty. They were more readily palpated and could be seen easily as small gray semitranslucent foci, closely simulating miliary tubercles. A bronchiole could be identified sometimes in the center of the nodule. From the third to the fifth or sixth day, the nodules were not so sharply outlined nor so translucent and firm as they became later. Not infrequently they were made prominent by the presence of hemorrhages in their substance, the color being thus changed from gray to bright red.

The microscopic picture in most of the cases was much the same; a wide-spread organizing bronchitis and bronchiolitis. (Figs. 70 and 71.) Under

FIG. 71- Higher magnification of bronchial wall shown in Figure 70. Fibroblasts are seen extending in a loose growth from the submucosa. One mitotic figure is present very low magnification one could see a patch of cellular tissue about almost every bronchus, especially the smaller (the tubercle-like nodules seen in the gross). The tissue was composed of young fibroblasts, a few capillaries, and many mononuclear wandering cells. In quite a number of the bronchi there was an exudate which was undergoing organization, with permanent obliteration of the lumen. Fibrin stains showed that there was often considerable old fibrin in the midst of the organizing areas.

Polynuclear leucocytes and fresh fibrin were not present in any quantity in the typical cases, which suggested that the infection (bronchitis or early bronchopneumonia of the first stage) to which these lesions undoubtedly were the sequel, had been successfully combated.



When death was delayed 3 to 10 days after exposure to phosgene, infection of the respiratory tract was the cardinal change found. The upper air passages remained practically unaffected except for a mild congestion, lout there was intense necrotizing infection of the bronchioles which not infrequently involved the surrounding alveoli and resulted in the formation of miliary abscesses. These were surrounded by small zones of hemorrhagic pneumonia. The intervening lung tissue showed alveoli filled with varying amounts of serum, fibrin and cells. If death was delayed more than four days, beginning organization of the exudate in the alveoli and bronchi was generally seen.

Animals which have apparently recovered from acute symptoms of gassing and which showed no signs of pneumonia, when killed 3 to 10 days after gassing, showed a widespread organizing bronchiolitis which clearly represented the sequel of the acute bronchial and peribronchial inflammatory reaction, so prominent in the acute period.


There were 177 dogs which survived gassing more than 10 days. Of these, 69 died and 108 were killed between the eleventh and one hundred and twenty-ninth day. Description will be made easier and clearer if the died" and 'killed" dogs are considered separately. It may be pointed out, however, that some of the animals that were killed looked sick and in poor condition and would have died in a few days had they been left alone. As might be expected, the lesions found in such animals are practically the same as in those that died.

The pathological findings in the dogs that died differed little from those observed in the 5 to 10 day animals already described. The essential and dominating feature in a majority of the cases was an infection of the respiratory tract. Reference to Table 54 shows that in 50 percent of the cases pneumonia of one type or another was present, and bronchitis in 75 percent. It can be stated safely that death was referable to respiratory infection, acute or chronic in at least 65 percent.

What, may be asked, was the cause of death in the remaining 35 percent?  In a few cases there was a chronic nephritis of the type not infrequently met with in dogs which may have been responsible. In the other 20 to 25 percent no cause of death was found. These animals were all poorly nourished and anemic. In other words, lesions sufficient to account for death in at least 20 percent of these animals were not demonstrated. It is possible, of course, that a careful study of the blood-forming organs or endocrine glands, not investigated in any of our cases, might have thrown some light on the question.


A majority of these were well nourished and healthy looking. Some were thin and sluggish, and a few were in bad shape and obviously about to die. The findings in these sick dogs have been referred to in the previous paragraph.

The gross changes in the healthy looking animals were not very striking. Outside the lungs, there was little worthy of note. The trachea and larger bronchi were quite normal.


The lungs were moderately collapsed, but the collapse was not uniform. Plate XII illustrates very well the picture often seen. There were ark pink atelectatic patches here and there; the rest of the lung tissue was more or less emphysematous. Tiny, firm nodules might be felt or seen on section, but these were not at all conspicuous. The bronchi were more prominent than normal. Their walls looked thickened and in the lumina there was an excess of mucus.

Microscopically, the picture was somewhat more varied. The emphysema and atelectasis were more pronounced than the gross appearance would indicate.

FIG. 72.- Organizing bronchiolitis in a dog killed 14 days after phosgene gassing. The lung grossly showed irregular patches of emphysema and ateleetasis, and tuberclelike nodules were felt throughout the lung

The alveoli were not otherwise altered. The bronchi showed marked changes. There was a distinct fibrous thickening of the walls of some of the medium-sized bronchi, and an infiltration by mononuclear wandering cells, most marked in the outer coat. In a much larger number of cases the only change found was in the small-sized bronchi. Here the lumen in places was completely occluded by a mass of granulation tissue with a zone similar to newly formed cellular connective tissue immediately about the bronchus. (Figs. 72 and 73.) This lesion, which was a perfect example of obliterative bronchiolitis. clearly represented a more advanced stage of the organizing bronchiolitis and pneumonia found in the (logs killed three to ten days after gassing. These changes


in the bronchi were quite sufficient to account for the persistence of the atelectasis and emphysema, which was seen to be directly proportional to the extent of the bronchial lesions. The susceptibility of these chronic dogs to pneumonia was also probably referable to the presence of such foci of infection in the bronchial wall. (See discussion of residual pulmonary lesions, p. 508.)


A comparative study was made in the American University laboratories of the effects of phosgene on various laboratory animals, including monkeys, guinea pigs, rats, rabbits, mice, dogs, and goats. It was found that the lesions produced in these animals were essentially the same. The observations made are well summarized in the following paragraph taken from the report of this work: 8

FIG. 73.- Higher magnification of two bronchi shown in Figure 72. The peribronchial thickening and the polypoid growths in the lumina are well shown

In the monkey and goat, for example, which represent the two extremes of susceptibility after exposure to the same concentration, lesions of the lung vary in degree, but not in character. The species variation, evidenced by the length of survival after gassing, depends in part upon the rate at which edema develops. On the other hand, some animals (monkey, guinea pig), the first to succumb to a given concentration, show less pulmonary edema than those that survive longer (dog, goat). This is evidence, as is brought out elsewhere just as clearly, that the edema is itself not the cause of death, but simply one manifestation of a more important underlying change.

While pulmonary edema develops more rapidly the more susceptible the species (monkey to goat) these animals-that is, the most susceptible,--show less edema than the more resistant ones.

This is an indication of the importance of the time interval in the production of the edema.

Among the investigations directed by the medical research committee of the British Army Medical Service, there is an excellent comprehensive study of the pathology of phosgene gassing in goats by Capt. J. Shaw Dunn. 8 Certain changes found, which have not been observed in dogs, make it worth while to review here in some detail Dunn's findings.


There is marked emphysema with irregular patches of atelectasis. Microscopically a wide-spread obliterative bronchiolitis is present.


In the first place it is pointed out that the use of goats in the study of poisoning by irritant gases affords certain advantages over most other laboratory animals.

1. The large size of the lungs permits of a much closer and more critical observation of naked-eye changes than is possible in the smaller laboratory animals.
2. The goat's lung is of a very definitely lobular construction, and in this respect is more fairly comparable with the lung of a healthy young man than those of smaller animals.
3. In goats the sequelae of gassing proceed almost invariably without septic complication, so that the phenomena observed are those attributable to the effect of the gas only.

All investigators may not agree that in the matter of size and lobulation the goat's lung has any definite advantage over the dog's lung, which is also well lobulated and quite large enough for satisfactory gross study. But certainly the relative immunity of the goat to respiratory infection made it possible to follow more easily the life history of the uncomplicated gas injury. On the other hand, it may be fairly argued that in respect to the tendency to septic complication the dog reacted much more like man than did the goat, and was therefore, for most purposes, a more suitable subject for experiment.

In general the changes found by Dunn in goats agreed with those already described in dogs, but there was one significant difference, namely, the presence of definite renal lesions.

The changes in the kidney were, in brief, a necrosis of the cells lining the convoluted tubules in the cortical labyrinth. The straight descending limbs of the convoluted tubules in the medullary rays, as well as the ascending limbs of Henle's loops, showed relatively little damage. Dunn suggests that the reason for the relative immunity of the straight tubules may depend on a different origin of their blood supply.

The renal necrosis was seen only in acute, fatal cases of gassing, and was found in only 49 out of 149 goats of this group. Similar changes were demonstrated also in acute deaths from chloropicrin.

As regards the conditions necessary to bring about renal necrosis, Dunn concluded that the lesion might be caused either by a high concentration for a short time, or by a lower concentration with prolonged exposures.f Artillery experiments showed that the lesion might be produced as effectively as in the gas chamber. In all cases, however, the exposure had to be lethal.

Dunn's observations on the reparative changes in the lungs agreed in most respects with those described for dogs. The rate of repair, however, was apparently more rapid in goats. Goats killed after the tenth day showed practically no trace of the gas injury beyond irregular capillary congestion and some thickening of the muscle fibers in infundibular and bronchiolar terminations, whereas in dogs, as has been pointed out, the lungs at this stage were full of fibroid nodules representing an organizing inflammatory reaction in the walls of the damaged bronchioles.


Phosgene produced in animals, as in man, a widespread injury of the parenchyma of the respiratory apparatus, followed by a series of reactive phenomena which might be complicated by septic infection. The initial damage, which
f Renal necrosis was found also in goats killed by chloropicrin, and in two animals, exposed respectively to diphenyl-chlorarsine and hydrocyanic acid.


involved the lining cells of the smaller bronchi, the alveolar epithelium, and possibly also the capillary walls of the septa, was fairly uniform in distribution, differing in this respect from the commoner bacterial injuries.

The results on the organism of this injury were. apparently entirely referable to local changes in the respiratory apparatus; evidences of a systemic intoxication, such as accompanies bacterial injury, were lacking.

The early inflammatory reaction consisted in a massive outpouring of fluid, in which considerable fibrin might be present. The cellular reaction, which consisted of polynuclear leucocytes and desquamated epithelium, was relatively slight. Obstruction of many of the smaller bronchial tubes was brought about by plugs of exudate and cell d6bris, and possibly also by irregular contractures of the muscle fibers of the bronchial wall. The result of such obstruction was partial or complete atelectasis of the area supplied by the occluded bronchus. A patchy compensatory emphysema occurred in the portions of the lung whose bronchial outlets remained patent.

In dogs and most other laboratory animals, bacterial infection frequently followed the gas injury, where death was delayed beyond 24 hours. In these cases pneumonia, generally lobular and necrotizing in character, resulted. Severe, but sublethal, exposures, not complicated by infection, led to a wide-spread organizing bronchiolitis, which simulated grossly miliary tuberculosis. These lesions tended to regress, ending in focal scars associated with more or less emphysema. Chronic infection of the bronchi with bronchiectasis was not an infrequent sequel of severe gassing.

No changes of any significance were found outside the respiratory tract, except in goats, where necrosis of renal tubules has been described following severe gas exposure.


P-884.-Male brindle bull terrier; weight 13.6 kilos. August 21, 1918, gassed for 30 minutes with phosgene, 83 parts per million. Found dead 8 a. m., August 22, approximately 18 hours after gassing.
Autopsy.- Body well nourished; weight 13.2 kilos; 0.4 kilo less than weight at time of gassing. Except for a well-marked congestion of the abdominal viscera, the positive findings were limited to the respiratory tract. The lungs weigh 35 grams, with a heart-lung ratio of 2.92. The lungs are greatly distended, deep purplish red, with scattered light-colored patches along the anterior margins. The tissue is boggy and noncrepitant throughout, owing, obviously, to the extreme edema. Upon section the lung tends to collapse somewhat with the escape of mulch fluid.
Microscopic findings.-The alveoli throughout the lung are more or less filled with granular precipitate (edema) with considerable fibrin. Very few free cells are seen, and those present are chiefly mnononuclears. The mucosa of the wall of the smaller bronchi is completely necrotic; that of the large bronchi and trachea is well preserved with cilia easily seen. The connective tissue about the larger bronchi and blood vessels is quite edematous with here and there a markedly distended lymph vessel.
Anatomical diagnosis.- Extreme edema and congestion of lungs; necrosis of bronchiolar epithehium.

NOTE.-The findings are typical of acute death from phosgene without septic complication.

P-734.-Female black and white cur; weight 5.7 kilos. July 10, 1918, gassed 30 minutes; concentration 69 parts per million. Found dead 4.30 a. in., July 11, 1918, 18 hours after gassing.
Autopsy.- Body poorly nourished; weight 5.45 kilos. Lungs weigh 225 grams. Heart-lung ratio, 3.94, or approximately three times the normal. They are voluminous and noncollapsible,


owing to the presence of extreme edema. Congestion is less striking than the edema. The amount of residual air in the lungs is very small. The trachea and bronchi contain much frothy fluid, but the mnucosa is normal looking, except for slight injection of vessels.
Microscopic findings.-Sections of the hlngs show much coagulated fluid and fibrin in alveoli, distention of capillaries, edema of perivascular tissue, and dilatation of lymph channel. There is, in addition, an early bronrchopneumonia, with an accumulation of polynuelear leucocytes in and about the bronchioles. Patches of greatly distended alveoli (acute emphysema) are seen near the surface of the lung. The epithelium of the smallest bronchi is quite necrotic, while that of the larger is well preserved. Sections of liver and kidney show no significant changes.
   Anatomical diagnosis.-Extreme edema and congestion of lungs; necrosis of bronchiolar epithelium; early bronchopneumonia.

NOTE.-The findings are typical of acute phosgene poisoning. The case shows very clearly the earliest stage of the pneumonic reaction, which is such a conspicuous feature in most of the later deaths.

P-661.- Male fox terrier; weight 10.55 kilos. July 1,1918, gassed 30 minutes; concentration 50 parts per million. Found dead 23 hours after gassing.
Autopsy.- Body well nourished; weight 9.53 kilos; 1.02 kilos less than at time of gassing. The lungs wveigh 440 grains; heart, 100 grams; heart-lung ratio, 4.40. The pleura and trachea are normal looking. The lungs are voluminous and show the usual mottling, owing to the congestion and patchy emphysema. The tissue is very doughy and noncrepitant, with an obvious extreme degree of edema. On section small indefinitely outlined gray patches are seen about the small bronchi, suggesting an early pneumonic reaction.
Microscopic findings.- The edema, congestion, and necrosis of the bronchiolar epithelium are the conspicuous features. In and about many of the bronchioles there is a polynuclear reaction, but this is distinctly focal. Small hemorrhages involving half a dozen or more alveoli are found here and there. The epithelium of the large bronchi and trachea is well preserved; that of the smallest air passages is necrotic.
Anatomical diagnosis.- Congestion and edema of lungs, with acute compensatory emphysema; necrosis of bronchiolar epithelium; early bronchopneumonia; miliary hemorrhages in lungs.

NOTE.- The lung picture is characteristic of acute phosgene poisoning with beginning respiratory infection.

P-785.-Young female terrier; weight 5.3 kilos. July 17, 1918, gassed 30 minutes; concentration 74 parts per million. Died July 20, 1918, three days after gassing.
Autopsy.- Body is fairly well nourished; weighs 5 kilos. or a loss of 0.5 kilo. Lungs weigh 300 grams and heart 84 grains, giving a heart-lung ratio of 3.57. The lungs fill the thorax and do not collapse. The tissue is boggy and airless, except for scattered emphysematous patches. The congestion is not so striking as in the more acute deaths, but the edema is extreme, as the weight index shows. The pleura and mediastinal tissues, as well as the lung, are quite edematous. The trachea and larger bronchi are apparently unaffected and examination of the abdominal viscera is likewise negative.
Microscopic findings.- The lung shows a patchy edema with an abundance of fibrin along the alveolar walls. There is a notable absence of a cellular reaction, except for occasional desquamated epithelium in the alveoli and a few lymphocytes and plasma cells in edematous perivascular and peribronchial tissues.
Anatomical diagnosis.- Edema and congestion of lungs and mediastinal tissues; necrosis of bronchiolar epithelium.

P-676.-Brown and white male cur; weight 15.7 kilos. July 2, 1918, gassed 8.27 to 8.57 a. m., with phosgene, 51 parts per million. Found dead 8 a. m., July 6. Degree of post-mortem change indicated that death had occurred about six hours previous, or approximately three and a half days after gassing.
Autopsy.- Body is poorly nonrrished but not emaciated; weight 14.16 kilos, or 1.54 kilos less than before gassing. Heart weighs 175 grams. It is filled with mixed red and white clot. There are no enidocardial hemorrhages. Except for-moderate distention of the chambers,


nothing noteworthy is found. The left pleural cavity contains a small quantity of blood-stained pus; while in the right cavity several hundred cubic centimeters of clear fluid is present, compressing the right lung. Lungs: Right weighs 140 grains; left 430 grains. The left lung is quite voluminous, deeply congested, but strikingly mottled owing to the presence of pale emphysematous patches throughout the organ. The upper lobes are moderately crepitant with scattered firm areas; the lower are firm and nodular throughout. On section, numerous patches of typical bronchopneumonia, tending to become confluent, are seen in the lower lobe, and similar, smaller an(l more scattered foci in the upper lobe. The right lung is small, firm and tough, and airless; the typical picture of pressure atelectasis. The trachea and larger bronchi show no gross change, except a slight congestion and some excess of mucus. The liver, kidneys, and spleen, which weigh, respectively, 725 grams, 58.6 grams, and 42 grams, are not grossly altered. Cultures of left lower lobe of lung show a Staphylococcus Aureus.
Microscopic findings.- The only significant findings are in the lungs. The picture varies in different portions. In the left upper lobe there is an early bronchiolitis with an interstitial reaction increasing the thickness of the septa. In the lower lobe, widespread pneumonic reaction is present. Atelectasis is the most marked change in the right lung. There is slight edema throughout. The mucosa of the medium-sized and large bronchi is well preserved; in the smaller air passages the layer of ciliated epithelium is lost. The deeper layer of cells shows evidence of active regeneration.
Anatomical diagnosis.-Bronchopneumonia (left) with beginninig organization; empyema (left); hydrothorax (right) with atelectasis of right lung; moderate edema and congestion; patchy compensatory emphysema.

NOTE.-This is a typical instance of delayed death from respiratory infection.

P-738.- Young male terrier; weight 5.9 kilos. July 10, 1918, gassed 30 minutes with phosgene; concentration 61 parts per million. Died, July 1.5, 1918, five days after gassing.
Autopsy.- Body is poorly nourished; weight 4.9 kilos; that is 1 kilo less than at time of gassing. The heart weighs 125 grains full of blood; 75 grains empty. It contains chicken-fat clot. Endocardium and myocardium normal looking. Luigs weigh 250 grains, giving a heart-lung ratio of 3.33. The lungs are very voluminous with a smooth, shiny pleura. They are moderately congested, very heavy, but not strikingly edematous. The tissue is firm and nodular throughout, with crepitation in patches only. On section the nodules are seen as poorly defined patches scattered through all lobes. There is no pus or other exudate in the larger bronchi or trachea. Lung cultures showed a Streptococcus hemolyticus.
Microscopic findings.-Throughout the lung there is considerable edematous fluid in the alveoli and the capillaries are everywhere congested. Fibrin is abundant. Special fibrin stains show dense blue masses of old fibrin adherent to the alveolar wall. In and about the smaller bronchi, there are many polyvutclear leucocytes. Elsewhere there are only a few cells in the alveolar spaces, and these are chiefly mononuclear. The lining epithelium of the smaller bronchi is entirely lost and the lumina are filled with cellular debris. Sections of liver, spleen, and kidney show moderate congestion but no other change.
Anatomical diagnosis.-Edema and congestion of lung; focal and diffuse interstitial bronchopneumonia.

NOTE.-The reaction of the lung in this case was strikingly proliferative, and suggests a response to a chemical rather than a bacterial injury. The presence of streptococci, however, indicates a double injury.

P-641.- Male brown bull; weight 12.4 kilos. June 28, 1918, gassed for 30 minutes with phosgene; 53 parts per million. Died, July 6, 1918, eight days after gassing.
Autopsy.- Body well nourished; weight 11.6 kilos, 0.8 kilo less than weight at time of gassing. The heart weighs 185 grams full of blood, 120 grams empty. There are no hemorrhages in the endocardium and no myocardial changes. The lungs weigh 475 grams with a heart-lung ratio of 3.95. The lungs are quite voluminous and on palpation are nodular in all lobes. The tissue is moderately congested and slightly edenmatous. There mire a few patches of acute emphysema. There are no hemorrhages. The nodules felt are seen on section as dry gray patches, fairly sharply defined. There is no exudate in the bronchi, apart from the expressed edema fluid.


Microscopic findings.- There is a widespread edema and congestion of the lungs. The fibrin is abundant and forms a dense layer on many of the alveolar walls. The septa are diffusely thickened owing to active proliferation of fibroblasts. In places there is a polynuclear cellular reaction in the smaller bronchi and adjacent alveoli, but much more conspicuous is the proliferative interstitial reaction. The epithelium of the smaller air passages is lost with here and there evidence of regeneration. In general the lining cells of the larger bronchi are intact, but in a few places a focal damage is seen.
Anatomical diagnosis.-Edema and congestion of lungs; necrosis and regeneration of bronchiolar epithelium; organizing interstitial pneumonia.

P-813.- Male collie. July 25, 1918, gassed 30 minutes; concentration, 87 parts per million. Recovered. Killed with strychnine August 6, 1918, 12 days after gassing. Condition when killed, good; looks well and is fairly lively.
Autopsy.- Body well nourished; weight, 13.6 kilos. No changes of any consequence are found except in the lungs, which show an organizing bronchiolitis of the type seen in P-668, though somewhat less marked; that is, the nodules are not so numerous. A few fresh hemorrhages are found in the pleura and lung. These obviously occurred in the death struggle.
Anatomical diagnosis.- Organizing bronchiolitis; recent pulmonary hemorrhages.

NOTE.-The findings are fairly typical of those in animals which were recovering from severe gassing. The lesions were obviously undergoing regression and are much less conspicuous than in dogs killed after several months, except in cases where a complicating pulmonary infection had occurred.

P-767.- Young female hound; weight, 6 kilos. July 15, 1918, gassed 30 minutes; concentration, 65 parts per million. Died, August 1, 1918, 17 days after gassing.
Autopsy.- Body poorly nourished; weight, 4.6 kilos. Except for a moderate degree of anemia of all tissues, significant changes are found only in the lungs. They weigh 175 grains, with a weight index of 3.43. There are widespread consolidated patches in all lobes, but most marked in the upper. On section the picture is a typical confluent bronchopneumonia with considerable edema and congestion.
Microscopic findings.-The bronchi form the centers of many- of the pneumonic patches. There is a proliferation of fibroblasts in the alveolar walls in places, but in general the inflammatory process looks quite recent.
Anatomical diagnosis.- Organizing bronchopneumonia; edema and congestion of lungs.

NOTE.-The case is an example of the development of a fatal pneumonia some time after the subsidence of the acute reaction following gassing.

P-668.-Female yellow hound. July 1, 1918, gassed 30 minutes; concentration, 55 parts per million. Recovered; regassed July 27; concentration, 74 parts per million. Killed with strychnia August 16, 1918, 36 days after first gassing, and 10 days after second gassing. Condition when killed, well nourished, lively, and apparently healthy, except for mange.
Autopsy.- Body weight, 14.5 kilos. Heart weighs 175 grams full, and 100 grams empty. The only significant positive findings are in the lungs. They weigh 275 grams, giving a heart-lung ratio of 2.75, and are moderately but irregularly collapsed. The anterior margins of the upper lobes are atelectatic. The pleura is thin and smooth, but here and there fresh subpleural hemorrhages are seen. While the lung tissue is crepitant throughout, numerous small firm nodules are felt. On section these are seen as indefinitely outlined red or gray foci. In many of the nodules recent hemorrhage is seen. There is no edema or congestion and no exudate in the bronchi.
Microscopic findings.-The number and uniform size of the nodules in the lungs, as they appear in the sections without magnification, suggest a miliary tuberculosis under the microscope, however, the lesion is seen to be an organizing bronchiolitis of the type shown in Figures 72 and 73. There is no leucocyte reaction and no fibrin or edema. The picture suggests a healing gas injury rather than a persistent bacterial infection.
Anatomical diagnosis.-Widespread organizing bronchiolitis ("pseudo-tuberculosis").

NOTE.- The extent of the pulmonary lesions in this ease was greater than in the majority of recovered dogs and was probably due to the double exposure.



Diphosgene (ClCOOCCl3), better known as "superpalite," is closely related chemically to phosgene (COCl2). It is therefore not surprising to find that the lesions produced in animals by the two gases are practically identical.Another closelv related compound, chloromethylchloroformate, better known as "palite," also produces effects which are not distinguishable from those of phosgene. Since the pathology of phosgene has been fully discussed, we shall give here only a brief sumrnary of the studies made by Winternitz and Wislocki,10 in the American Universitv laboratories, leaving the interested reader to consult the original report for details.

Studies were made of changes in 35 gassed dogs, of which 21 succumbed within 3 days after gassing, and 9 from 3 to 14 days. Five dogs recovered and were killed 2 weeks to 3 months after gassing.

As with phosgene gassing, there was a latent period after exposure in which no deaths occurred. The earliest death was 8 hours after gassing, while the majority of animals succumbed between 12 and 36 hours.

In the acute deaths, pulmonary edema and congestion, and patchy emphysema were the most striking gross features. Histological studies showed that, as in phosgene, the trachea and larger bronchi were spared, the chief seat of injury being the distal portion of the respiratory apparatus--that is, bronchioles and alveoli.

Dogs that survived the acute stage were very prone to respiratory infection, and deaths occurring after two days were practically all due to pneumonia.

A study of the lungs of recovered dogs showed the same residual lesions--bronchiolar scarring, patchy emphysema, chronic bronchitis--that were observed in phosgene dogs.


While chloropicrin belongs to the respiratory irritant group of gases, it has certain properties which set it apart from the other members of this group. In the first place, it is not a gas but a liquid with a fairly high boiling point (112° C.). Furthermore, in contrast to phosgene and chlorine, it is very stable and nonabsorbable, and while in its deleterious action on tissues decomposition undoubtedly takes place, the nature of the change is not known.

Direct application of the liquid to the skin produces severe, deep burns, and a drop on the cornea results in ulceration. Inhalation of the gas leads to respiratory and circulatory disturbances, closely resembling those associated with chlorine and phosgene gassing, already described. Tendency to nausea and vomiting, apparently from the irritation of the stomach by the swallowed gas is more common than with the other war gases, though not constant, except where high concentrations are used.

A systematic study of the pathology of chloropicrin gassing was made in the Yale laboratories 11 through cooperation with other workers, who were investigating other phases of the gas action. The following report is based on the observations thus obtained.


There were, altogether, 120 gassed dogs, upon which autopsies were performed. Histological as well as gross studies were made in all cases.


TABLE 55.- Dogs gassed with chloropicrin

Table 55 shows the number of deaths in successive days after gassing, and in Chart XXVII there is a curve based on these figures. It is seen that, as with corresponding lethal exposures to chlorine and phosgene, the first 24 hours after exposure was the critical period, the largest number of deaths

CHART XXVII.- Duration of life after chloropicrin gassing.

occurring during this time. The first part of the curve is not like the curves for chlorine and phosgene, in that in the former the peak is reached in the first 12 hours, instead of the second 12 hours, as is the case with the latter. However, there is not the secondary rise on the third and fourth days, noted in the chlorine series.

No figures, as to the number of animals which survived gassing, are given, but it may be stated that in general a fairly large percentage of animals survived exposure to concentrations of 60 to 80 parts per million whereas a half hour's exposure to somewhat higher concentrations (125 to 150 parts per million) gave a mortality of more than 50 percent.


The changes found at autopsy, as might be expected, were quite similar in many respects to those produced by chlorine, phosgene, and other members of the respirator irritant group.


Since the experimental pathology of chlorine and phosgene has been discussed in some detail in the following description, frequent comparisons will be drawn of the differences between the lesions produced by those gases and chloropicrin.

We may conveniently divide the discussion into the findings in animals dying (1) in the acute stage and (2) in the subacute and chronic periods.


We have applied the expression "acute stage" somewhat arbitrarily to the first 24 to 48 hours after gassing, during which time the signs and symptoms were chiefly those of edema of the lungs, without definite evidence of infection.


The following protocol will illustrate very well the usual findings at this period:

Young Airedale, female; weight 11.1 kilos; gassed September 26, 1918, 30 minutes; concentration 1.035 mgm. per liter.
Typical symptoms during and after gassing: Lacrymation, retching and vomiting; depression. Later, viscid frothy discharge from mouth and nose, rapid and labored breathing, restlessness. Death, 15 hours after gassing.
Autopsy.- Three hours after death. Body is well nourished, but flanks and abdomen are shrunken. Weight is 700 grams less than at time of gassing. Body is lax; there are beginning post-mortem changes.
Abdomen.- Negative, except for engorgement of veins and congestion of liver. The congestion disappears as vessels are cut, the organ assuming normal color, with usual distinct lobulation.
Thorax.-The cardiac area is large; the pericardial sac is tense, being stretched by the distended heart. Lungs are bulky, but do not tend to overlap the heart. There is a slight excess of fluid in pleural cavities and pericardium. Tissues about the great vessels are slightly blood stained.
Heart.-Weighs, full of blood, 130 grams; empty, 83 grams. Heart-body ratio is approximately 0.008, that is, about normal. The right side is more distended than the left. Chambers are filled with red clot. The heart valves are normal. Beneath the endocardium of the left ventricle there are three or four hemorrhagic patches 1 to 3 mm. wide, and extending 0.5 to 1 cm. along the crests of the muscular ridges. The heart is otherwise normal.
Lungs.-Weight, 320 grams; volume, 330 c. c. The lungs are voluminous but not extremely so. Pleura is smooth, but looks thick and slightly opaque (edematous), giving the suggestion of a film over the lung. The lungs are unusually colored, being a dusky, bluish red or pink, distinctly cyanotic, with a semitranslucent quality over all. Here and there are light whitish patches, but these are neither large nor numerous (as in phosgene poisoning). In such patches there is crepitation, but practically everywhere else the lung is doughy and airless. On section, clear fluid and blood pour out like water from a squeezed sponge. The tissue is red and semitranslucent, with occasional lighter, air-containing patches. Near the margins of the upper lobes there are a few small emphysematous areas. The bronchi, like the lung proper, are full of fluid, and in the large branches there is some froth. The mucosa is somewhat reddened, and that of the smallest branches is rather opaque. About each there is a zone of edema, which is also conspicuous about the larger blood vessels. The trachea is full of sticky froth. Its mucosa is slightly reddened but is otherwise normal.
Examination of the remaining organs, liver, spleen, kidneys, adrenals, gastrointestinal tract, brain, is negative.
Microscopic findings.-There are no noteworthy changes except in the respiratory tract. The tracheal epithelium is practically everywhere intact, but in places the superficial cells are somewhat shrunken and distorted and have lost their cilia; a few are desquamated. In the largest bronchi a similar condition is seen; but as one passes downward into the medium-sized cartilage-containing tubes, the injury is far more serious. The superficial cells are quite necrotic, and the entire layer is partially loosened from the wall. (Fig. 74.) In the bronchioles and atria there is not only death of the lining cells but necrosis of the wall itself.


FIG. 74.- Necrosis of bronchial epithelium and subepithelial edema in acute death from chloropicrin gassing


The lung tissue shows practically everywhere a complete filling of the alveoli with coagulated edematous fluid, which is quite rich in albumin. (Fig. 75.) A few air bubbles are seen, but these are more prominent in the bronchial fluid. Desquamated alveolar cells are fairly numerous. There is some fibrin free in the alveoli, and covering the septa everywhere, like vines on a lattice, are dense strands.
Special stains show that the fibrin threads permeate the alveolar walls interrupting the capillary bed, as has been demonstrated in phosgene and chlorine poisoning. This permeation is most marked in and about the walls of the bronchioles where the damage to the tissue is greatest.
Anatomical diagnosis.- Extreme edema and congestion of lungs; necrosis of bronchial epithelium and bronchiolar walls; dilatation of heart; passive congestion of abdominal viscera.

FIG. 75.- Widespread edema of lung associated with acute death from chloropicrin. Note occasional clear spaces (air bubbles) in some of the alveoli

The findings in the case just described are quite typical of acute chloropicrin death. There was little difference in the picture whether the animal died 4 hours or 24 hours after gassing, except that in those living longer an early inflammatory reaction was generally found. More will be said of this in connection with the "delayed deaths."

There was in all of these early deaths an overwhelming edema of the lungs, which constituted the most striking feature of the autopsy findings. The degree of edema, as judged by the lung-heart and lung-body weight ratios, varied considerably with individual animals, but the average for dogs dying in the first, second, or third 6-hour period was practically the same. This


might suggest that death ensued in each particular case when a certain degree of edema was reached. But an analysis of the figures obtained from dogs which were apparently recovering, killed 18 to 48 hours after gassing, shows that among there the degree of edema, based on the proportionate weight of the lungs, heart, and body, was as great as in the animals that died. In other words, the presence of any definite quantity of fluid in the lung did not seem to be the cause of death. This idea has been substantiated experimentally by Winternitz and Smith, 12 in their studies upon pulmonary irrigation. This question will be considered in a later part of this chapter dealing with the more important phenomena of the gassed state.


The proportion of animals dying on any one day after the first 48 hours was relatively small (See Table 55), but the total number of these "delayed deaths" was considerable in any large series of experiments. The cause of death was in almost every instance a superimposed respiratory infection. This might begin within a few hours after gassing, but generally did not become widespread for several days. In some instances the infection ran a chronic course, killing only after weeks or even months, in which cases a suppurative bronchitis with more or less organizing pneumonia was generally found.

Dogs dying 2 to 10 days after gassing usually showed a purulent bronchitis and bronchopneumonia. (Fig. 76.) Several lobes were almost always affected. Occasionally, patches of consolidation were found throughout the entire lung. The upper lobes were more often involved than the lower. This may be accounted for either by the greater dependency of the upper lobes in four-footed animals, or by the poorer drainage due to the more abrupt branching of the bronchi. There was a somewhat greater tendency to abscess formation and extension of the infection to the pleura than after phosgene gassing. Organization of the pneumonic exudate occurred frequently.

TABLE 56.-Dogs gassed with chloropicrin- percent showing pneumonia

It will not be necessary to describe the gross and microscopic findings in these cases since they were practically the same as in chlorine and phosgene poisoning, which have been fully discussed in the preceding pages.


Twenty-five animals, which had survived exposure to a lethal g concentration of chloropicrin, were killed at different periods after gassing. (See Table 55 for detailed figures.) These exhibited very clearly the nature of the reparative processes which follow the gas injury.

g The term "lethal" is applied here to concentrations which, in a given period kill the majority of exposed dogs.


The edema began to regress after about two days, but at least a week was required for its complete disappearance. It may be mentioned in this connection that the quantity of fluid present in the lungs in some of the two and three day dogs, in which all untoward symptoms had disappeared, was found to be as great as in the dogs which had died. There was, however, a greater quantity of residual air in the lungs, as shown bv a comparison of the weights and volumes.

Fibrin in the alveolar spaces and walls was removed rather slowly, old strings and plugs of it being demonstrated sometimes seven or eight days after gassing.

FIG. 76.-Acute bronchitis and bronchopneumonia causing death three days after exposure to chloropicrin. The lung tissue separating the pneumonic patches is markedly edematous

Where the bronchiolar walls had been seriously damaged, and this happened quite regularly after exposure to high concentrations, and active proliferation of fibroblasts began about the dead area as early as three days after gassing, and the bronchial cavity was soon filled with granulation tissue. The final picture was that of an obliterative bronchiolitis, as described in chlorine and phosgene gassing.

Regeneration of the bronchial and alveolar epithelium proceeded quite rapidly. It was best seen in dogs which had been exposed to sublethal Concentrations of the gas. where only the superficial layer of cells had been killed.


In sublethal exposures, there were present quite regularly curious structures in the alveolar spaces, which under low magnification looked like giant cells. Under greater magnification they had the form of capillaries, although one found in all only disintegrated red blood cells and fibrin, and no fresh blood. One was inclined, at first, to interpret these structures as foreign body cells, which were formed about masses of cell debris and fibrin, but a more careful study showed that they were attached to the alveolar wall and were probably sections of bulging capillary tufts which had become thrombosed.

TABLE 57.- Dogs gassed with chloropicrin-degree of edema of lungs


Chloropicrin, like chlorine and some of the other gases of the respiratory irritant group, injured the epithelium of the entire respiratory tract, but all portions of the tract were not equally affected. The trachea and largest bronchi, though irritated, suffered only slight and transient injury. The medium-sized and small bronchi were most affected. There was a uniform, widespread damage of the alveolar walls, which, however, was not severe enough to lead to necrosis. The alveoli were apparently nowhere protected by constriction of the bronchi.

An overwhelming edema of the lungs rapidly followed exposure to lethal concentrations of the gas. In extreme cases practically every alveolus was filled with fluid, so that at autopsy the weight and volume of the lungs (expressed in grams and cubic centimeters) approximated one another. In addition to the fluid in the lung itself, there was also marked edema of the mediastinal tissues and pleura, which was even more striking than in phosgene and chlorine


gassing. The edema fluid contained fibrin in places, and a great deal of fibrin was found in the alveolar walls. It was especially abundant in dogs that had lived at least 24 hours.

Partial or complete occlusion of the smaller bronchi by inflammatory exudate or masses of necrotic desquamated cells led to focal emphysema or atelectasis, but this was not such a striking feature at autopsy as in death from some of the other respiratory irritant gases (phosgene, superpalite).

Infection of the lungs, with the development of a widespread bronchitis and bronchopneumonia, was seen in a large percentage of the severely gassed animals which did not die in the first few hours after gassing. Abscess formation, pleurisy-fibrinous or purulent-and organizing pneumonia were common complications. In recovered animals there was a regeneration of the epithelium of the bronchi and alveoli, and organization of the necrotic bronchiolar wall, with scar formation (obliterative bronchiolitis). Focal atelectatic and emphysematous patches remained as permanent gross evidence of the gas injury.

A study of "recovering" animals killed at different periods indicated that the cause of death in the early stage-that is, before infection had become well established-was not due to the edema per se but probably to obstruction of the blood flow through the lungs caused by extensive deposition of fibrin in the alveolar walls. The increased viscosity of the blood from the loss of fluid into the lungs, emphasized by Underhill,5 was no doubt also a very important factor. Likewise some of the "delayed deaths" were to be attributed to this obstruction in the pulmonary circulation, but the great majority were obviously due to an infection of the lungs, bronchi or pleura.


C. P. 34.- Brown and white male mongrel; weight 13.1 kilos. Gassed, August, 23,1918; 30 minutes exposure to ehloropierin; 169.3 parts per million. Greatly excited during exposure, juminping around and licking nose. Toward the end there was vomiting and urination. Half hour later, pulse 60, respirations 16, temperature 39.2 °. Eight hours later pulse 132, respirations 68, and temperature 38.6 ° ; hemoglobin 229 percent. Death occurred 10 hours after exposure.
Autopsy.- Body weight 12.5 kilos (a loss of 0.6 kilo). The lungs weigh 375 grams, the heart 125 grams, giving a heart-lung ratio of 3. The lungs are mottled and voluminous, and quite doughy. They are airless in the entire lower lobes and the greater part of the upper lobes. Crepitation is felt only along the margins of the upper lobes. There are a few subpleural hemorrhages.
Microscopic findings.- The lungs show a widespread edema, the fluid in the alveoli being particularly rich in albumin, as indicated by the deeply staining coagulum. (See Fig. 75.) There is practically no cellular reaction. The mucosa of the bronchi is quite necrotic and sloughing; that of the trachea is damaged with partial desquamation in places, but no reaction except marked vascular congestion.
Anatomical diagnoses.- Congestion and edema of lungs; necrosis of bronchial epithelium.

NOTE.-The case illustrates well the intermediary position of the lesions of chloropicrin when compared with those of phosgene and chlorine. The trachea, for example, was damaged much more than in gassing by phosgene, but much less than was seen after fatal exposures to chlorine.

C. P. 35.- Female coach dog; weight 15.4 kilos. Gassed. August 23, 1918; 30 minutes exposure to chloropicrin; 164.9 parts per million. Dog took gas well, showing very little disturbance. Before exposure, pulse 66, respirations 36, and temperature 38.3°. Six hours after gassing, pulse 124, respirations 24, and temperature 38; hemoglobin 213 percent. The dog was found dead 12 hours after gassing.


Autopsy.- Body weight was 14.1, a loss of 1.3 kilos. The lungs weigh 490 grams; heart weighs 175 grams; heart-lung ratio 2.8. The lungs are dark purple, with a slight mottling in the upper lobes. There is edema throughout, with extreme congestion. Only a few patches of emphysema are present and these are limited to the upper lobes. There are no hemorrhages or pneumonic foci. The trachea and bronchi show congested mucosae, but no membrane formation.
Microscopic findings.- There is extreme congestion and edema in the lungs, with no pneumonia. Changes are practically the same as in C. P. 34, described above.

C. P. 93.-Boston bull; weight 7.15 kilos. September 9, 1921, gassed 30 minutes with chloropicrin; 0.697 mgm. per liter. During exposure vomits, licks nose and chops, and keeps eyes closed. Found dead on morning of September 11, two days after gassing.
Autopsy.-Body weighs 5.5 kilos (loss of 1.65 kilos after gassing). Heart weighs 55 grams, lungs 310 grams; heart-lung ratio 5.63. Lungs are heavy and voluminous, with well-marked edema and congestion, in addition to firm, confluent pneumonic areas which are felt in all lobes. Trachea looks unchanged except for slight congestion.
Microscopic findings.- A typical bronchopneumonia is found such as is shown in Figure 76. The bronchi are filled with purulent exudate and necrotic epithelium.
Anatomical diagnoses.- Edema and congestion of lungs; necrosis of bronchial epithelium; bronchopneumornia.

NOTE.-The findings are typical of those generally seen in delayed deaths.The respiratory infection dominates the picture.

C. P. 6.- Black and brown male hound; weight 14 kilos. Gassed, August 16, 1918; 30 minutes exposure to chloropicrin, 103.5 parts per million. There was retching and vomiting during exposure, with salivation and discharge from the nose. One hour later, pulse 68, respirations 38, and temperature 38.5 ° . Condition did not appear serious, and three days later dog was sent to the Brady Laboratory showing no symptoms. Dog was killed with chloroform.
Autopsy.- Body weighs 12.5 kilos, which is 1.5 kilos less than at time of gassing. Lesions are found only in the lungs. They weigh together 190 grams, with a heart-lung ratio of 1.79. The lungs are somewhat mottled and voluminous. All lobes are mare or less doughy and on section exude fluid. A few scattered hemorrhages are seen.
Microscopic findings.- Only subacute changes are found. There is active regeneration of the lining epithelium of the smaller bronchi and bronchioles. The bronchial walls are quite cellular, from the presence of both fibroblasts and mononuclear wandering cells. Nearly every bronchus shows hemorrhagic extravasations, as is very well shown in Plates XIII and XIV. In the alveoli about the small bronchi are many large phagocytic cells, and occasional capillary-like structures such as is seen in Figures 77 and 78. On the whole, the fibroblastic reaction is not as marked as is generally seen after chlorine and phosgene gassing.
Anatomical diagnoses .- Organizing brornchiolitis with hemorrhages; degeneration of bronchial epithelium.

NOTE.-The hemorrhages in this case were evidently terminal occurring during the death struggle. The other lesions were probably regressive and would have gone on to healing.


Chlorine injures by combining directly with the cytoplasm of exposed cells, or through the formation of hydrochloric acid, which, in turn, acts on the tissues. Phosgene, COCl 2 coming in contact with water, rapidly decomposes, with the liberation of hydrochloric acid, HCl, which constitutes the toxic agent. Chloropicrin is very stable in vitro, and it has not yet been determined just how it reacts with tissues to injure them; but it seems likely that, as in the case of phosgene, its toxicity is referable to the chlorine part of its molecule, which is in some way split off as the gas reaches the tissue.


It would appear, therefore, that with each of the three gases the directly injurious agent is the same-chlorine-and that their pathologic effects should be very similar. The description of the lesions produced in dogs by the three gases, as given in the preceding pages, shows that this assumption is correct. There were, however, certain points of difference, which are sufficiently clearcut to enable an experienced observer to say from an examination of the organs of a gassed animal which of the three gases had been used.


Chlorine damaged, and in high concentration entirely destroyed, the epithelial lining of the upper portion of the respiratory tract-trachea, large

FIG. 77.- Lung of "recovered" dog killed four days after chloropicrin gassing. In alveoli surrounding the bronchioles there are structures resembling giant cells. As shown in Figure 78, the structures are composed of fused mononuclear cells inclosing hits of old fibrin and degenerated red cells

and medium-sized bronchi. Although the injury may have extended to the distal alveoli as well, causing focal areas of necrosis in the lung (see figs. 51 and 52) and desquamation of the alveolar epithelium, the most severe injury was suffered by the trachea and bronchi.

Phosgene, on the other hand, spared the trachea and larger bronchi, but destroyed the epithelial lining of the smaller bronchi and bronchioles. Even the outer coats of the smallest air tubes showed evidence of serious injury, and the alveolar walls were everywhere damaged though rarely necrotized.



Note particularly the flattened character of the regenerating bronchial epithelium. A mitotic figure is seen on the left.


Chloropicrin occupied an intermediary position in its action on the respiratory epithelium. The lining cells of the trachea and the very large bronchi were definitely injured in places, as shown by irregularities in the ciliated surface or loss of cilia and occasional desquamation of the superficial layer of cells. But there was nowhere seen the rapid and complete coagulation of the entire mucous surface such as chlorine produced. The medium-sized and smaller bronchi, on the other hand, suffered very severe damage, which was even more marked than with phosgene. There was often complete disintegration of the walls of the bronchioles. The changes in the alveolar walls were practically the same as those produced by phosgene.

This difference in the behavior of the three gases toward the several portions of the respiratory tract is not easily explained, but must be related in some way to the fact that chlorine required no preliminary decomposition for its action, thus hitting hardest the first tissues with which it came in contact; whereas phosgene, and probably chloropicrin also, must be broken up, a process which may take place in themoist air of the smaller bronchi, or in the cytoplasm of the lining epithelium while being, absorbed, or, as Hoover 4 has suggested in the case of phosgene, only after absorption. The place and rate of absorption and decomposition would thus determine the site and degree of injury.


Each of the three gases produced a high grade of edema of the lungs which developed with great rapidity after exposure. Its development was more rapid with chlorine and chloropicrin than with phosgene, where correspond toxic concentrations were used. In the two former a maximum degree of edema was reached in less than 12 hours; whereas after phosgene the peak was reached in about 18 hours. But the time varied greatly with individual animals.

 FIG. 78.- Higher magnification of an alveolus from Figure 77, showing the structure of phagocytic giant cells. Note their resemblance to capillaries

The actual quantity of fluid poured out into the lungs, as judged by the lung-heart and lung-body ratios, was slightly greater with phosgene than with chloropicrin, although with the latter the fluid filled the lungs more completely (no figures are available for chlorine). This point will be taken up again in connection with the volume of the lungs. In both chlorine and chloropicrin gassing there was a very striking edema of the mediastinal tissues and the pleura; this was not as conspicuous with phosgene.

Congestion of the lungs, although present in all, was much more pronounced with phosgene than with the other two gases, and constituted one of the distinguishing gross features. Phosgene lungs were almost regularly dark bluish red or purple, with whitish emphysematous patches; chloropicrin lungs, on the other hand, as a rule, were lighter in color, the basic hitue being a bluish pink.



The development of focal emphysema was seen with all three gases, but phosgene easily ranked first in the number and extent of the acutely distended foci of lung tissue. A comparison of the illustrations of gross lesions emphasizes the differences. Figures on the weights and volume of chloropicrin and phosgene lungs supply more exact data, particularly as regards the total amount of residual air.

In 18 phosgene dogs dying within 24 hours after gassing, the average volume of the lungs was 579 c. c.; weight 459 grams. In 30 chloropicrin animals the figures were 413 c.c. and 353 grams. The average weight of the dogs in the two series was approximately the same, 12 kilos. Phosgene lungs were seen to be greater in both volume and weight, but the difference between the average weight and volume, which represented roughly the amount of residual air confined chiefly in emphysematous patches, was 129 in the phosgene series, while in the ehloropicrin series it was less than half this amount (see Table 57 for chloropicrin figures). The contrast in the gross appearance of the lungs was often much more striking than these figures would suggest. In many instances the chloropicrin lungs were uniformly doughy and airless, and the figures of the weight and volume were practically the same, or differed by only 10 to 25 points. With phosgene no such close approximation of the weight and volume was observed, the smallest difference noted being 50, this being in a small dog, weighing only 5.5 kilos, where the volume of the lung was 325 c.c. and weight 275 grams.


An inflammatory reaction was seen in the lungs of almost every animal that died more than 24 hours after gassing with chlorine, phosgene, or chloropicrin, and in a large proportion of those dying earlier. In this particular, there was practically no difference in the three gases. But the percentage of animals which succumbed to the infection after passing through the acute edema stage appeared to be somewhat higher with chlorine than with the other two gases. The secondary rise in the death curve (see Chart XXIV) shows this to be so. In view of the greater injury to the trachea and bronchi produced by chlorine, which allowed not only the freer entry of pathogenic bacteria into the lungs, but also led to a necrotizing tracheitis and bronchitis as well, the particularly large number of delayed deaths from infection was not surprising. As might be expected, chlor opicrin came second to chlorine in producing a condition favorable to infection. Indeed, in the matter of infections of the pleural cavity-fibrinous pleurisy and empyema-it appeared to stand first, but the series of cases was small, and the percentages were therefore less conclusive.

The late changes in the lungs-that is, the lesions found in animals which had died or had been killed ten days to several months after gassing-were practically the same for all three gases: focal emphysema and atelectasis, chronic bronchitis, generally of the obliterative type, and occasional examples of bronchiectasis. An active chronic infection in and about the bronchi, with patches of organizing pneumonia, was seen not infrequently. On the whole these chronic changes were most pronounced after chlorine gassing, but the character of the lesions was the same in all, and similar pictures have been described for other gases of the respiratory irritant group.


In other words, it was only in the acute period that it was possible to distinguish between the effects of these gases, and even then it could be done only by an experienced observer who was familiar with the variations in the lesions produced by each gas.


A comparative study of the pathology of chloropicrin, chlorine, and phosgene shows that chloropicrin in its action on the respiratory tract, occupied an intermediate position between chlorine and phosgene. It damaged the trachea and larger bronchi less than chlorine, but more than phosgene. In its action on the bronchioles and alveoli, it resembled phosgene very closely, and in several other respects the lesions were more like those of chlorine. The gross and microscopic differences in the acute effects of the three gases on dogs were sufficiently clear to enable an experienced observer to determine by autopsy which gas had been used. It should be possible to make practical application of this knowledge on the battlefield in the identification of the gas being used by the enemy.


The experimental pathology of three compounds of this group, hydrocyanic acid, cyanogen chloride, antl cvanogen bromide, was investigated at the American University laboratories by Winternitz, Finney, and Wislocki.13 The following is a summary of their report.

The rapid and fatal action of hydrocyanic acid and certain of its salts was referable to its property of interfering with tissue oxidation, so that, as Geppert 14 stated many years ago, "there is an internal suffocation of the organs." The involvement of vital centers in the central nervous system accounted for the sudden death when a lethal dose was administered. It was not surprising that with such rapid action, little or no distinctive anatomical change should be produced. The blood was bright red, and coagulation was delayed, resulting in a scarlet instead of the usual bluish lividity in the dependent parts. If potassium cyanide reached the stomach undecoinposed, it formed, with the hemoglobin, a striking red or blue cyanmethemoglobin compound. The mucous membrane was soapy, slippery, covered with blood-tinged mucus, and quite transparent at the crests of the folds. This change was brought about by the alkaline action of the potassium, and was regarded as characteristic. It was claimed by Kobert 15 that increased pressure in the ventricles of the brain might occur in the more delaved action of the poisoning. The right heart was usually distended, the left empty. Post-mortem digestion set in rapidly, especially in the liver. The lungs might be edematous, and the urine might contain blood and sugar.

With hydrocyanic acid there was no essential delayed action, and experimental results (Marshall 16 ) indicate that if an animal did not succumb during exposure to the gas, it recovered, and there were no anatomical changes found when the animal was sacrificed. It follows, therefore, that changes which occurred in animals subjected to chlorine and bromine compounds of cyanogen were probably, in large part at least, the result of the action of the halogen radicals.



The action of this gas was similar to that of hydrocyanic acid in that a lethal exposure resulted in death within a few minutes, apparently from paralysis of the respiratory center. Marshall and Miller,16 in a large series of experiments, observed no delayed deaths, and they stated that if an animal survived 15 to 20 minutes after removal from the gas chamber it would recover. In this respect the gas differed quite strikingly from chlorine, phosgene, and chloropicrin, which did not cause immediate death except in extremely high concentrations.

The anatomical changes in cyanogen chloride poisoning were studied in two groups of dogs: (1) Those that succumbed during or immediately after exposure, and (2) recovered animals killed at varying intervals up to two months after gassing. The former showed a moderate degree of pulmonary edema and congestion, associated with an early inflammatory reaction in the bronchioles. The edema was inconspicuous with minimal lethal doses, but it was definitely augmented as the concentration of the gas was increased. The recovered animals showed a catarrhal bronchiolitis associated with slight localized vesicular emphysema and atelectasis, and rarely, small foci of bronchopneumonia.

These findings show that in addition to a cyanogen action, there was a definite, though mild, injury to the respiratory tract, which was undoubtedly referable to the halogen radical.


The action of this gas was somewhat different from that of the chlorine derivative. While animals might succumb suddenly from short exposure, death some hours or even several days after exposure was more common.

The lesions in acutely fatal cases (death within 24 hours) resembled closely those caused by chlorine. The entire respiratory apparatus suffered, particularly the trachea and bronchi. The lung showed the same picture of extreme edema and congestion as was seen in gassing by other respiratory irritants.

Animals dying after 48 hours generally showed well-marked respiratory infection; tracheitis, bronchitis, and bronchopneumonia, associated with more or less edema and congestion. As in the more acute deaths, the changes found were practically indistinguishable from those caused by chlorine.h The marked involvement of the, upper respiratory tract differentiated the condition from phosgene poisoning.

Animals which recovered from a fairly severe exposure showed quite regularly residual lesions in the lungs of the nature of a chronic bronchiolitis, organizing or suppurative with more or less emphysema (Fig. 79).


The chemical properties an(l physiological action of arsine have been discussed elsewhere. Since, however, the anatomical changes produced by lethal exposures to the gas are intimately related to the action of the poison on the blood, it may be well to restate in this connection the salient facts regarding this action. Arsine is absorbed through the lungs and passes into the blood
h Microscopic sections of the organs of these animals have not been examined . It would be interesting to see whether the lungs show focal hyaline necroses of the septa such as are found in fatal chlorine gassing.


stream apparently unchanged. There it enters into combination with the hemoglobin of the erythrocytes, forming a compound which gives a brownish color to the blood. After several hours, the affected blood cells rupture (hemolysis), setting free hemoglobin, which is excreted by the kidneys, appearing in the urine both as methemoglobin and oxyhemoglobin, and by the liver which converts it into bile. An excess of free hemoglobin is found in the blood and is absorbed by many of the tissues, particularly those in direct contact with the blood stream, such as the intima of blood vessels. It is important to note that this striking hemolytic action is dependent on the presence of oxygen or some

FIG. 79.- Organizing bronchiolitis in a dog killed two months after exposure to cyanogen bromide. The character of the cells taking part in the organization is well shown. The bronchiole has been partially relined by an epithelium of irregular form but mostly of the flattened type

oxidizing substance, and that a period of several hours is necessary for the completion of the reaction. This latent period has been observed both in vivo (prejaundiced stage) and in vitro.

As might be expected, one of the most serious effects of such a hemolytic agent is the rapid development of an anemia of severe grade. The blood changes in a dog receiving a sublethal exposure are shown very strikingly in the rapid fall in the number of red blood cells, with a gradual restoration to normal as the animal recovered. There was an associated leucocytosis, presumably


a toxic reaction. Wislocki 17 states that the hemoglobin in severe cases may fall to 20 percent or even lower, and that the erythrocyte count not uncommonly sinks to one or two million within 72 hours. Blood films show basophilic stippling, poikilocytosis, and anisocytosis. The appearance in the circulation of numerous normoblasts and occasional megaloblasts indicates an active attempt at regeneration on the part of the bone marrow.


The following summary of the pathological anatomy of arsine poisoning is taken from Mackenzie's 18 report of observations at the American University.

Arsine produced a pathological picture which had practically nothing in common with the changes characterizing the gases known as lung irritants. Locally, there was no irritant action, animals showing no distress during exposure; nor did the bronchi or lungs, through which the gas was readily absorbed, show any anatomical changes suggesting that there was direct action upon any of the tissues of the respiratory system. As a rule the lungs were grossly quite normal in appearance, except for areas of atelectasis at the bases and along the margins of the lower lobes. These, like the small amount of edema in the alveoli, seen in microscopic sections, and the occasional appearance of effusions in the pleural cavities, probably appeared just before death, coincident with failure of the circulation and extreme oxygen want. It was upon the blood, and particularly upon the red blood cells, that the chief action of arsine was exerted.

At autopsy, the skin, mucous membranes, intima of blood vessels, and the serous layer of the gut were stained a dusky red color. This staining of the tissues obscured somewhat the jaundice which was constantly present. In the sclerae, liver, and fatty tissues, however, it was very apparent. The post-mortem clots in the heart had a uniform, almost black, appearance without any of the chicken-fat quality. The liver was swollen and jaundiced and the gall-bladder full of dark, viscid bile. The enormous production of bile was shown by the presence of large amounts throughout the whole alimentary tract from the esophagus to the rectum. The spleen was enlarged and soft; often its anterior tip extended to the midline of the abdomen. It was deep bluish in color, and on section the pulp was dark, soft and smeary; the follicles indistinct or invisible. In addition to the staining of the serosa and the abnormally large amount of bile, the intestines sometimes showed blood-stained contents. No ulcerations or other gross lesions, however, were seen. Pancreas and adrenals showed only the dusky red discoloration.

The kidneys presented a remarkable appearance. They were swollen and soft. The capsule stripped off easily, leaving a smooth surface which, in the more severe poisonings, was almost black in color. When the action was less intense, the surface showed scattered patches of blue-black color on a back-ground of deep red. On section the cortex was wide and bulging, and the striations, though less distinct than in a normal kidney, were regular and straight. Both cortical and pyramidal portions showed the dark blue-black discoloration. The bladder contained brownish black urine, which, in thin layers, had a reddish tinge. In the bladder there was also found a precipitate of black amorphous material which yielded a strongly positive test for arsenic; a watery extract of it gave the spectrum of methemoglobin. The bone marrow was dark, grayish-red, and firm.



Lung.-A small amount of coagulated fluid was found in the alveoli. This was probably not the result of irritant action by the gas, but merely a terminal event, comparable to the pulmonary edema so frequently found at autopsy in cases of pernicious anemia.

Heart.-Nothing more than pallor of the muscle fibers was found.

Liver.-The liver cells were swollen and frequently contained an abnormal quantity of fat. In the bile canaliculi and in the lymphatics of the portal spaces dark greenish-brown droplets of bile were seen.

Spleen.- The follicles were large, but the cells seemed less closely crowded together than normally. The pulp contained relatively few cells. It was filled with formless material, staining, in hematoxylin and eosin preparations, a pale pink or pale greenish-yellow color. Evidently this material was hemolyzed blood and the debris of laked corpuscles. A few nucleated red cells and plasma cells were found in the pulp and a scattering of large mononuclear phagocytes frequently laden with pigment. The histological picture gave the impression that the splenic enlargement was due to the filter action of the pulp removing from the circulation the fragments of destroyed corpuscles.

Stomach, intestines, pancreas, and adrenals.-Nothing abnormal. Degeneration of the cells of the adrenal medulla, described for poisoning by other arsenic compounds, was not seen.

Kidneys.- The capsular spaces and lumen of the tubules were filled with hemolyzed blood and often, also, hemoglobin crystals. Sometimes one saw the straight tubules apparently plugged by a mass of crystals. In the cytoplasm of the epithelial cells lining the straight tubules, and to a less extent in the epithelium of the convoluted tubules, accumulations of small dark-brown crystals were frequently found. In two of the animals was seen degeneration of the epithelium lining the convoluted tubules. It is quite possible that this was the result of obstruction by crystals in the collecting tubule. One dog, which survived the gassing nine days, showed evidences of epithelial regeneration. The process in the kidney, therefore, appeared to be as follows: Free hemoglobin or a combination of hemoglobin and arsenic was removed from the circulation by the glomeruli and tubules. If the concentration of hemoglobin was sufficiently high, crystallization took place, and the crystals becoming impacted in the lumen of the collecting tubules, caused obstruction. This resulted in a rise of pressure within the obstructed tubule and consequent degeneration of the epithelium. The gross discoloration of the kidneys was due in part to the deposition of crystals in the cytoplasm of the kidney cells, but chiefly to the filling of the tubules with methemoglobin, and the vessels of the kidney with the brown product of hemolysis by arsenic oxide. An obvious therapeutic indication from these findings is the administration of large amounts of fluid in order to keep the concentration of hemoglobin in the kidney at the lowest possible point.

Bone marrow.- There was no evidence of any direct action on the marrow. With the development of anemia, however, the stimulus of oxygen want caused a striking hyperplasia of the marrow. The increased activity was shown by myeloblastic as well as erythroblastic cells.

The following case illustrates very well the effects of arsine on the blood cells. Dog exposed 30 minutes; concentration 0.53 mgm. per liter. Six


hours after gassing there was hematuria, which continued four days. Other symptoms were diarrhea, tarry stools, marked weakness, and depression. Eventually recovery was complete. Table 58 shows the quantitative changes in red blood cells and leucocytes. Note the rapid fall in erythrocytes, which reached their lowest figure on the fifth day, after which there was a gradual restoration to normal. There was at first a fall in the leucocytes, then a marked rise, and finally a gradual decline. Normoblasts and megaloblasts appeared in the circulation in the period of active regeneration.

TABLE 58.- Blood changes following gassing with arsine


Wislocki 17 makes the following note on pathological changes in animals which survive "'lethal" exposure and are killed after partial or complete recovery.

After four to six days the jaundice, which is so characteristic of the second stage, fades. The hemoglobinuria ceases; the viscera lose their icteric tint, but for some time possess an abnormal pallor, attributable to the anemia from loss of blood.

The red cell counts and hemoglobin estimations gradually rise. Evidence of the stimulation of the hemopoietic function is seen in the presence of immature erythroblasts in the peripheral circulation. Microscopically, the bone marrow is found slightly hyperplastic, showing numerous erythroblasts in all stages of development.

Liver and spleen show no trace of their former condition beyond, under the microscope, the presence of pigment granules in mononuclear phagocytes.

The kidneys do not appear to have been severely damaged, for every trace of the methemoglobin disappears, and the injured tubules are restored to normal. Deposits of hemosiderin are encountered for some time.


The pathological changes found in arsine poisoning were referable directly or indirectly to the lytic action of the gas on the red blood-cells. An acute anemia resulted, with a staining of the tissues by the freed hemoglobin. Jaundice developed rapidly.

Microscopically, methemoglobin was found in amorphous and crystalline form in the lumina of the kidney tubules and in the renal epithelium. Bile pigments were deposited in practically all organs and tissues.

In recovered animals there was active and rapid regeneration of the blood, with bone marrow hyperlasia. The blood pigments were gradually removed from the tissue apparently by phagocytosis.



The number of toxic organic compounds of arsenic is quite large, as pointed out in the discussion of the chemistry of this group of substances. But relatively few organic arsenicals were utilized in the World War. Diphenylchlorarsine, diphenylcyanarsine, and chlormethylarsine are mentioned by Clark and Pappenheimer 19 as being those most employed.

Diphenylchlorarsine and diphenylcyanarsine are solids which in finely divided form produce an extremely irritating and penetrating smoke, causing sneezing and lacrymation (blue cross shells). Bromethylarsine and chlormethylarsine are liquids which in vaporized form injure the respiratory tract, eyes, and skin in very much the same way as does mustard gas.

In addition to these four compounds, a number of closely related substances including ethyl-, methyl-, and phenyldichlorarsine, ethylarsine, methyl-diodoarsine, cacodylcyanide and chloride, diphenyliodoarsine, and diphenylarsine oxide were investigated in the American University laboratories, but thorough anatomical and histological studies were made only in the case of methyl-, ethyl-, and phenyldichlorarsine. Preliminary studies, however, with various members of the group, indicated that while they might vary considerably in toxicity, the lesions produced by lethal exposures were essentially the same in all.

Investigations carried out by Dunn 20 for the medical research committee of the British Army Medical Service also showed that the effects of the several arsenicals studied, diphenylchlorarsine, diphenylarsenious oxide, and phenylarsenious oxide, are identical, and that the effect produced in different species of animals (goats, dogs, guinea pigs, cats, monkeys, and rabbits) are much the same.


These compounds were very thoroughly investigated at the American University. Dogs were subjected to various concentrations of each of the vaporized substances in a gassing chamber. As with other toxic gases studied, it was found that a certain percentage of the exposed animals died shortly after gassing, whereas some succumbed only after several days, and others completely recovered. The numbers falling into each of these groups, particularly the acute deaths and recoveries, varied, as might be expected, with the degree of concentration of the gas and the length of exposure. The lesions associated with death in the first stage, that is within the first two days, were most distinctive of the gas injury, since in the later deaths, the gross and microscopic picture in the respiratory tract, where the damage was greatest, was regularly obscured by the reaction to a superimposed infection. The lesions produced by the three compounds differed in degree rather than in kind and may therefore be discussed together.

The following summary of Finney's observations 21 on the effects of methyldichlorarsine will serve as a basis for all.


After exposure to high concentrations, death might occur within 4 hours, but rarely occurred before 18 to 24 hours. While in the gassing chamber the dogs showed signs of marked irritation and often became excited. There was profuse lacrymation, salivation, and nasal discharge, accompanied by sneezing and


often by retching and vomiting. After removal from the chamber, the animals were depressed, had no appetite, and began to exhibit more or less respiratory distress. Vomiting was frequent and diarrhea might be present. After a few hours, wheezing and rales might be heard, and the rate of respiration was increased. The pulse became rapid and feeble as the respiratory signs were augmented, and the animal became semicomatose and died in collapse. The pathological findings in these acute animals were as follows:

The conjunctivae were reddened and congested; there was often a frothy discharge from nose and mouth; the skin of the groins where sparsely covered by hair frequently showed red patches, extending throughout the thickness of the corium.

On opening the thorax, the lungs were voluminous, nearly filling the chest. All lobes usually shared equally in the increase in size and other changes. The pleural surfaces were moist, without the presence of free fluid in the pleural cavities. The surfaces were smooth, bright red in color, with more or less mottling over the surfaces, caused by alternating groups of air-containing and serum containing alveoli. On removal, the lungs were felt to be heavy and the lobes retained their shape well, although hanging in a sodden mass from the hilus. On section, the cut surfaces dripped frothy pink fluid mixed with blood. The color was not uniform, but areas of darker red appeared here and there throughout the lungs. These areas were often firmer in consistency than the more normal spongy tissue, and gave the impression of jelly which had begun to set.The anterior tips of the middle and upper lobes most frequently showed these dark red areas.

The bronchi were usually not conspicuous on the cut surface of the lung, though they were sometimes filled with tenacious fibrinous plugs. The trachea, however, showed a striking false membrane throughout its entire length. The membrane was quite thick and edematous, sometimes quite white in color, in contrast to the walls of the trachea, which were congested and often intensely red. More frequently the membrane was stained pink by the frothy fluid which poured up from the lungs. In the earliest stages, the membrane could usually be stripped entire from the surface beneath; later on it lost its tough elastic quality, became yellowish, softer, heavier, and finally, purulent and necrotic. The surface beneath was red and raw, and occasionally small bleeding points could be seen. The larynx, up to the vocal cords, showed, a continuation of the membrane; above this point, the congestion and edema of the mucosa was still present, but the membrane was lacking. The formation of this membrane took place quickly and was found in dogs dying as early as four hours after exposure. In many of these extremely acute dogs there was no membrane but, instead, congestion and some edema of the mucosa. In dogs dying at the end of the second day or later the membrane was replaced by thick, purulent exudate.

The heart was dilated and filled with dark red, almost black, post-mortem clot, particularly on the right side. The pericardial fluid was not increased in amount and was not bloody.

The abdominal viscera might show moderate degrees of congestion, but this was never a striking feature. The liver was firm and dark red. The spleen varied considerably and was usually slightly swollen, presenting a smooth, soft surface; on section its Malpighian bodies were not conspicuous. The


kidneys presented a smooth surface, a (lark red cortex, with usually a lighter pyramid, and no marked abnormalities. The adrenals looked entirely normal on the surface; but on section, the medulla was usually markedly congested and occasionally pin-point hemorrhages could be seen at the line of juncture of cortex and medulla. The intestines showed nothing of especial interest; the mucous surfaces were not congested. The urine found in the bladder had its normal yellow color.

Microscopically the only changes of importance were in the respiratory tract. The membrane in the trachea was found to consist of a fibrin network with large spaces full of edematous fluid. The epithelial lining of the trachea had been lifted off and in some cases could be seen as a ragged line floating out through the edematous fibrin. In the early stages, there was but little polymorphonuclear infiltration; but this became more marked in a short time, and the membrane might stain a dusky purple, because of the presence of pycnotic nuclei of leucocytes. The submucosa also showed some leucocytic infiltration, but this never became extreme, and many of the leucocytes were mononuclear; the edema of the submucosa was also of a moderate degree. Sometimes the inflammatory process extended beyond the cartilaginous rings. Later on the membrane became a dense mass of necrotic purple-staining material, which lay upon the submucosa, with no intervening epithelial layer. It was noticed in many cases that while the epithelium was completely denuded over the greater part of the circumference of the trachea, it was still intact over the posterior portion; that part which was not surrounded by cartilage.

In the lungs the microscopic changes were principally those associated with the congestion and edema seen in the gross. The edema was rarely of even distribution throughout the section, but groups of alveoli were seen filled with pink-staining homogeneous material adjacent to other groups of alveoli over-distended with air. (Fig. 80.) This corresponded to the mottled appearance presented by the surfaces of the lungs. The capillaries of the alveolar walls, particularly in the edematous areas, were tortuous and filled with red blood cells. The alveolar epithelium seemed to have suffered little, and was in most places intact; where edema was most marked, there wits more or less desquamation and swollen round cells were found floating in the fluid. In places much fibrin could be seen in the edematous alveoli. The bronchi showed a continuation of the tracheal membrane (fig. 81), but the bronchioles usually presented an intact epithelium. Frequently there was marked peribronchial and perivascular edema. (Fig. 82.)

Microscopic changes in the other organs were of minor degree, and consisted of slight swelling and granulation of cells in the liver and kidneys, and of congestion of the vessels of the adrenal medulla with occasional minute hemorrhages.


The dogs of this group showed more variation than those of the first group. Those which died in two or three days showed a persistence of the acute conditions described above, upon which was superimposed a beginning pneumonia. Those which die(l toward the end of the two-week period showed little trace of the original conditions, but had extensive pneumonia of various types.

It appeared from a study of the microscopical sections that the pneumonia could develop in two ways. First, the extensive edema of the acute


stage might become diffusely infiltrated with polymorphonuclear leucocytes, and the edema might change to a cellular exudate over large areas of lung, bearing little direct anatomical relationship to the bronchi. Second, the smaller bronchi and bronchioles could become ringed with leucocytes, their epithelium degenerated, their walls infiltrated, and each bronchiole the center of a small focus or nodule of pneumonic consolidation.

In the first case the consolidated area may have occupied all of a lobe or several lobes, and the lungs may have retained a heavy, wet consistency. In the second case the areas involved were practically always the anterior tips of the middle and upper lobes, while the posterior portions of these lobes and the whole of the lower lobes presented an air-containing and emphysematous condition, sometimes with slight congestion, sometimes with none.

FIG. 80.- Widespread edema of lung in acute death from dichlorethylarsine. The clear round spaces indicate the presence of air bubbles

The pneumonias were of different types: Some were soft, wet, and purulent, with the bronchi containing streams of purulent material; others were firm, gray, solid pneumonia; still others were red, with many small discrete gray nodules standing out on the cut surface.

In dogs which lived for some time after the development of pneumonia, abscesses were occasionally found, with fibrinopurulent pleurisy. (Plate XV.) The acute conjunctivitis of the early stage subsided and rarely developed into a purulent discharge. Ulceration of the cornea was found in only one or two isloated cases. (Fig. 83.) The skin lesions also subsided without ulceration, blistering, or sloughing.


Some dogs which survived for three weeks or more after exposure died with acute pneumonia of only a few days' duration. Others died with extensive purulent pleurisy or empyema. There seemed to be no definite proof that


Only one lobe is consolidated; others are relatively emphysematous.


these animals developed this condition as a result of exposure to the gas, as some animals came to autopsy with the same lesions before exposure to any gas.

Those dogs which survived for long periods and were sacrified for autopsy showed no gross anatomical lesions. Microscopically, no changes were found in the lungs or bronchioles (in the trachea there was sometimes evidence that the epithelium had not recovered its normal thickness or structure).


From the above description it is clear that methyldichlorarsine produced an escharotic action on the tubular portion of the respiratory tract, including larynx, trachea, and the main bronchi. and that there was marked though less

FIG. 81- Necrotization of the bronchial lining associated with acute death from dichlorethylarsine. A few air bubbles are seen in the exudate filling the lumen

severe damage to the branches of the bronchial tree and the alveolar walls.The resulting lesions were a membranous laryngitis (figs. 84 and 85), tracheitis and bronchitis (fig. 86), edema and congestion of the lungs. The smaller bronchi were often plugged by masses of exfoliated epithelium dislodged downwards from trachea and large bronchi. A focal emphysema and atelectasis were found as secondary phenomena, and hemorrhages into both alveoli and interstitial tissue about the blood vessels and bronchi at times constituted an added striking feature. Small hemorrhages were much less commonly found in the adrenal and liver.

Among the animals which survived the acute period, a certain number, as indicated above, died after 2 to 10 days or even longer. In practically all of these, the chief cause of death was found to be a widespread infection of the respiratory tract, generally a suppurative bronchitis and bronchopneumonia;


abcess formation and more rarely a suppurative pleurisy were present occasionally. In other words, the picture was not different from that associated with similar delayed deaths after mustard or chlorine gassing. Finney's studies of recovered animals killed considerable periods after gassing with methyldichlorarsine showed neither grossly nor microscopically any chronic or healed lesions. 21 The significance of this observation is discussed on pages 483, 508 in connection with the residual lesions of the various gases of the respiratory irritant group.

The effects of ethyldichlorarsine and phenyldichlorarsine Finney found, in general, to be practically the same as those produced by the methyl compound just described. A few minor differences were noted. For example, the ethyl compound seemed to be more potent in the production of hemorrhages than

FIG. 82.- Marked perivascular edema and dilatation of lymphatics in acute dichlorethylarsine gassing

the other two, whereas phenyldichlorarsine produced the most severe tracheal injury. It is questionable, however, if even these minor points of difference may not be referable to differences in toxicity.

A comparison of the effects of these organic arsenic compounds with gases more widely used in the war--chlorine, phosgene, chloropicrin, and mustard--shows that the lesions of the former most closely resembled those produced by mustard gas. Indeed, it would have been quite difficult, either grossly or microscopically, to distinguish the tracheal lesions in the two cases. Finney pointed out, however, that the organic arsenicals of this group did not damage the distal portion of the bronchial tree, as mustard gas did, and that the eye effects were different. The arsenicals produced an acute irritation of the eyes, with much discomfort, as evidenced by the behavior of the animals, but the injury was slight and the conjunctival reaction transient. Mustard gas, on the other hand, caused much less irritation--that is, subjective disturbance--but far more serious damage to the ocular tissues.


With arsine (AsH3) which acts primarily as a blood destroyer or hemolytic agent, the group of organic arsenicals apparently has nothing in common from the standpoint of anatomical effect. It is also noteworthy that injury to the kidney, such as Pearce and Brown22 found, was produced by a number of the organic arsenicals which they tested, was not observed in dogs gassed by any of the four compounds discussed here, although owing to the different mode of administration (subcutaneous and intravenous in the one and pulmonary in the other) the tests are perhaps not properly comparable.

It would appear, therefore, that the toxic property in methyl-, ethyl-, and phenyldichlorarsine, as in mustard gas, is linked up with the chlorine radical, rather than the arsenical group. Certainly the changes in the tissues give support to this view.

FIG. 83.- Ulceration of cornea following exposure to dichlorethylarsine. Perforation of the cornea has taken p1ace with infection of the anterior chamber


It is obvious from the detailed description of the pathological changes produced by toxic exposure to the various pulmonary irritant gases, that certain phenomena-pulmonary edema, circulatory disturbances, focal emphysema and atelectasis, complicating respiratory infection and hemorrhages, and residual lesions in recovered animals-were common to all. These changes have already been fully described. What is proposed here is simply the discussion of the significance of some of these phenomena from the anatomical standpoint.


Two phases of the question of pulmonary edema merit discussion: (1)The cause and mechanism of production; (2) the importance of the condition as a cause of death after gassing.


In respect to the cause of the edema, Edkins and Tweedy, 2 in a very thorough discussion of this fundamental question, set down the following main possibilities:

(1) It may be simply a response of increased lymph production to some appropriate local stimulation.
(2) It may be the result of damage to the alveolar epithelium and the capillary walls, such that increased permeability occurs as far as the plasma and its constituents are concerned, though not permitting diapedesis of erythrocytes. In severe poisoning, however, this may even occur.
(3) The damage done to the alveolar walls may lead to the formation of new chemical substances which osmotically induce a passage of water from the capillaries. It is perhaps even possible that the irritant substances themselves, or their chemical products, may be of minor importance in this respect.

FIG. 84.- Blister formation in epiglottis in a dog dying acutely from exposure to dichlorethylarsine. The epidermis is severely damaged and there is a widespread inflammatory reaction in the tissues beneath

(4) It may be the result of such disorganization of the walls of the vessels or spaces containing lymph that definite leakage is established.

These investigators mentioned also the possibility of defective lymphatic drainage as an important factor. The problem may be stated in somewhat simpler terms. Is the fluid a mere transudate which pours out through walls made pervious by direct injury or through disturbed circulation? Or does its coagulation represent a true inflammatory reaction comparable to that seen in the skin or other solid tissues, resulting from bacterial or chemical injuries? The inclination was to accept the latter view, namely, that the edema is a part of an inflammatory reaction and therefore essentially purposeful, although it seems impossible to exclude purely mechanical factors. Certainly the alveolar epithelium is damaged, and possibly also the adjacent capillary wall, though this is not readily demonstrable. But an injury that


would permit such a free outflow of fluid might be expected also to allow the exodus of red blood cells, whereas the active diapedesis of red corpuscles is exceptional. Furthermore, the presence of fluid in the interstitial tissue of the mediastinum and tracheal walls is more easily explained on an inflammatory basis. In supporting the idea of the inflammatory nature of the edema, it is granted that the mechanism of such a reaction is by no means settled. It would be of interest to review this problem here, but it would take one too far afield.i It may be added that there is accord with the views of Edkins and Tweedy, as expressed in the following paragraph: 2

We are further led to the view that the edema is a consequence of excessive lymph production and that this production is the adapted response which the lung makes to the stimulus afforded by the irritant gas, or the changes in the tissues such irritant gases immediately

FIG. 85.- A higher magnification of one of the blisters shown in Figure 84. The fluid of the bleb is rich in fibrin which stains deeply

induce. The accumulation, though its advantage in maintaining or restoring the lung tissue may be great, has the disadvantage that it may fatally interfere with the gaseous exchange. However necessary its appearance may be to combat the poison, when in excess, it is as desirable to get rid of it as in the case of pus in an abscess.
i Auer (Proc. Soc. Exper. Biol. and Med., 1917-18, XV, 106) observed a localized edema in cats following the inhalation of dimethylsulphate. Generally one lobe or part of a lobe was involved. Auer expressed the belief that this localized fluid accumulation was due to the combination of a partial or complete stenosis of a bronchus or bronchiole along with an inspiratory dyspnea. "Under these conditions," he says, "each alveolus which is in connection with a stenosed bronchus or bronchiole will act like a miniature dry cup during inspiration because the pressure in these alveoli will decrease as the intrathoracic pressure decreases during each inspiration, for little or no air enters through the stenosed air passage. Therefore, during each inspiration which is slow, labored, and prolonged in the gassed cat, the capillaries of the alveolar walls are subjected to an aspirating action which facilitates or initiates the passage of a transudate into the alveolar spaces. The production of this transudate is probably also aided by a local damage of the alveolar epithelium which the war gas produces. Practically these observations may be of some value. In gassed soldiers all inspiratory dyspnea should ameliorated as mu(h as possible by tracheotomy and artificial respiration if necessary." While edema of the lungs produced by the commoner respiratory irritant gases is diffuse rather than localized, it is possible that the mechanism suggested by Auer, namely, bronchial constriction or obstruction, plus inspiratory dyspnea, may be a factor in bringing about the rapid and extensive fluid accumulation in the air sacs.


In other words, the edema is to be regarded as a purposeful reaction, which may, and if excessive undoubtedly does, lead to a certain degree of pulmonary embarrassment.

The importance of edema, as a cause of death, is discussed in the physiological section of this volume, but is considered here from the anatomical standpoint.

That the edema of the lungs is the immediate or direct cause of death from gassing by chlorine or phosgene seems to be taken for granted in most of the clinical reports of human cases.

Physiological studies likewise have tended to confirm the view that fluid in the pulmonary alveoli interferes with gaseous interchange and that when this interference with respiration passes a certain critical point, the patient dies of asphyxia, or, as popularly stated, drowns in his own fluid. A full discussion

FIG. 86.- Complete necrosis of the mucosa of a large bronchus resulting from dichlorethylarsine gassing. There is a marked inflammatory reaction throughout the wall of the bronchus

of the mechanism of death in gas poisoning is given elsewhere in this volume. (See p. 354.) There are presented here certain anatomic data which, it is believed, tend to show that undue emphasis has been placed on pulmonary edema as, per se, a cause of death. These observations, together with the results of some experiments upon what may be termed an artificial edema produced by filling the lungs of a normal dog with an isotonic solution have led not only to the questioning of the importance of pulmonary edema as a cause of death, but to the conclusion that edema of the lungs in general is merely an indication of some underlying disturbance, and is rarely, if ever, solely responsible for the death of the patient or animal.

The observations may be briefly summarized:


(1) Great variations are seen in the degree of pulmonary edema associated with death from acute gassing by any of the pulmonary irritant gases. With very high concentrations of gas or long exposures, death may occur quickly with little edema; on the other hand, sublethal exposures may give rise to marked edema with recovery. Animals recovering from gassing often show, when killed, a degree of edema greater than others which have succumbed. Furthermore, in certain species, such as rats and guinea pigs, the edema is inconspicuous, although the animals are quite as susceptible to gassing as other species in which edema is a marked feature.
(2) There is no relation between the degree of pulmonary edema and the concentration of the blood, which, as Underhill has shown, is a fair index of the seriousness of the gassed state.
(3) The lungs of a normal animal may be filled in great part with isotonic salt solution, thus producing an artificial edema, comparable in degree with that produced by gassing, without resulting in serious embarrassment of the respiration or circulation. Several of the above propositions require some elaboration or discussion.


Marked differences in degree of edema may be readily judged by inspection and palpation, but for careful comparative study, a more accurate quantitative method, capable of expression in figures, is necessary. Such a method, based on the comparative weights of the lungs and empty heart, has been suggested and used by Barcroft in investigations carried on under the auspices of the British Medical Research Committee. 23 As used in the present investigations the method was as follows:

The lungs were weighed with the trachea attached, cut to uniform length, and clamped to prevent the escape of edema fluid. The heart was also trimmed uniformly and completely emptied. The normal ratio of the lung weight to the heart weight was obtained by taking an average of 16 normal animals. This normal ratio was found to be 1:1.30, or simply 1.3. (Barcroft's higher figure, 1.5, is probably the result of a difference in the method of trimming of the organs.) In gassed dogs the lung weight was divided by the heart weight, and this quotient by the normal ratio, 1.30. The resulting figure was termed the "edema index." It represented the percentage increase in lung weight over the normal.

This method was objected to by Eyster, 24 who insisted that the dried weight method was much more reliable. There were two of these dried weight methods: One, used by some of the French investigators, in which a typical slice of the lung was weighed wet and then after drying, and the second, the method used by Eyster and his assistants, in which the entire lung was dried, the proportion between the wet and the dried specimen indicating the degree of edema. The first method was obviously open to large errors. The second method had the serious drawback of making it impossible to study the lung grossly or microscopically. The latter method was compared with the more simple lung-heart ratio in a series of six dogs, and it was found that the two methods gave results approximately the same (compare columns 9 and 13 in Table 59).


TABLE 59.- Relation of the edema of lung and the concentration of blood in gassed animals.

In using the lung-heart ratio method of estimating the amount of edema fluid present, it was found that the edema index in a series of dogs gassed under similar conditions varied within relatively wide limits. For example, among 50 dogs dying after exposure to phosgene, the edema index ranged from 1.73 to 4.60, and in another series, gassed with chloropicrin, the extremes were 1.65 and 4.22.

In order to throw further light on the question of the significance of the degree of edema, the following experiment was made: Eight dogs were gassed with phosgene (concentration 80 to 90 mgm. per liter for 30 minutes). Four of the dogs died in from 10 to 15 hours. The remaining four were killed by chloroform. It was found that the average edema index of the four dogs that died was practically the same as that of the four that were killed. j It was found also that many of the dogs which passed successfully through the critical forty-eight hour period and were then placed in the "recovered" group showed, if killed at this stage, a high edema index, often exceeding that of the dogs which had succumbed. It may be stated incidentally that these "recovered" dogs showed no symptoms other than occasional coughing and a slight sluggishness.

Still further evidence of the subsidiary part of edema as the cause of death after inhalation of irritating gases was found in the comparative effects of a gas such as phosgene on animals of different species.

A series of experiments was performed in which a number of different kinds of animals were exposed in the same chamber for thirty minutes to a concentration of 0.27 mgm. per liter of phosgene. The time of survival varied as indicated in Table 60.
j Ricker, G. (Samml. Kim. Vort. Inner. Med., Leipsig, 1919, Nos. 256-260. 727), in a study of the lung changes in cats gassed with phosgene, determined the degree of edema at varying intervals after gassing, using the weight method. Examination of his tabulated findings shows that one animal that was killed after 48 hours showed a greater edema than two others that died between 9 and 24 hours after gassing; a result that corresponds with our observations.


TABLE 60.- Animals gassed with phosgene

Species: Time of survival
Monkey...............................................3 hours 30 minutes.
Guinea pig...........................................4 hours 30 minutes.
Rat.......................................................5 hours.
Rabbit..................................................1 hours 30 minutes.
Mouse...................................................Killed after 12 hours.
Dog.......................................................Killed after 12 hours.
Goat......................................................Killed after 12 hours.

The lesions produced in these animals by inhalation of phosgene were essentially alike. In the monkey and goat, for example, which represented the two extremes of susceptibility after exposure to the same concentration, lesions of the lung varied in degree but not in character. The species variation, evidenced by the length of the survival after gassing, in animals which had been killed or had died, could be expressed in part by the rate of development of the pulmonary edema. On the other hand, with some animals (monkey, guinea pig), the first to succumb to a given concentration showed less pulmonary edema than those that survived longer (dog, goat). This is evidence that the edema was in itself not the cause of death but simply one manifestation of a more important underlying change.

While the pulmonary edema developed more rapidly the more susceptible the species (Table 60) the most susceptible showed at death less edema than the more resistant ones. This is an indication of the importance of the time interval in the production of the edema.


Underhill 5 has found that dogs exposed to phosgene, chlorine, or chloropicrin showed after a few hours (the time varies with different gases and with individual animals) a well-marked increase in the concentration of the blood. Similar changes in the blood of gassed soldiers have been demonstrated repeatedly. The formed elements of the blood as well as the inorganic salts shared in the change. Inorganic salts, however, did not follow the same course as the erythrocytes. The result was a marked increase in blood viscosity. Underhill and his assistants used this blood change as an index of the condition of the gassed animals, and upon it have worked out a method of therapy, the essentials of which are bleeding and subsequent dilution of the residual blood with isotonic salt solution. In applying this method of treatment, which, it may be stated, has definitely reduced the mortality among experimentally gassed dogs, it has been assumed that the concentration of the blood was due to the escape of blood serum into the lungs, and that, therefore, the increased viscosity of the blood could be taken as a rough index of the degree of pulmonary edema.

In order to determine whether or not these two phenomena, blood concentration and pulmonary edema, were directly related, the following experiment was carried out. Twelve dogs were gassed, nine with phosgene and three with chloropicrin, the duration of exposure and concentration being such as would be fatal to a majority of dogs exposed. The dogs were killed with chloroform about 10 hours after the exposure to the gas; that is, as soon as the majority began to show serious symptoms. A red blood cell count was made before gassing and again just before the animal was chloroformed. since it has


been shown that this is a reliable method for estimating the degree of blood concentration. The degree of pulmonary edema found was determined by the method described above. The figures for the increase in blood concentration and the edema index are recorded in Table 59 with other data. The results are also graphically shown in Chart XXVIII. It is seen that under the condition of these experiments no parallelism existed between the amount of fluid present in the lung and the degree of concentration of the peripheral blood. It is noteworthy that in one case in which there was a well-marked edema of the lung, an actual reduction of blood concentration was found. k This experiment

CHART XXVIII.- Comparison of the degree of pulmonary edema and concentration of the blood in gassed animals. Cross-hatched columns represent percentage of increase in lung weight and are arranged in order. Solid black columns represent percentage of increase in red blood corpuscles. This chart clearly shows that there is no relationship between the two.

does not indicate that the loss of fluid from the blood may not have taken place by way of the lungs and the mouth, but, in our opinion, it does show conclusively that the change in the blood does not serve as an indicator of the amount of fluid present in the lung at any given moment. It suggests, further, that a therapy guided by the viscosity of the blood can not be assumed to have any influence on the pulmonary edema, and that the beneficial results obtained by such therapy are probably in no way referable to a change in the fluid content of the lung, which of itself is of secondary importance, as will be emphasized in the following paragraph.
k There was dilution of the blood, as shown by Underhill, which preceded its concentration with phosgene poisoning. These changes of blood concentration may vary somewhat in duration. etc.. and explain the charted findings above.



Winternitz and Smith 12 have shown that the lung is much less susceptible to the introduction of fluid than has been generally supposed. Repeated experiments have demonstrated that the lungs can be flooded through the bronchi with isotonic salt solution and that this process of irrigation can continue for at least two hours without causing any evident harmful changes in bodily well-being, or any subsequent serious microscopic lesions in the lung tissue. By means of the use of colored solutions it has been shown that the fluid introduced actually passes throughout the lung, bronchi, bronchioles and alveoli, and does not simply flow back through the trachea without entering the lung.

Examination of the lungs of dogs killed immediately after irrigation showed a marked "edema," comparable in respect to the amount of actual fluid in the lungs to that of the gassed animals. There was, however, little or no damage to the respiratory parenchyma, or the pulmonary circulation, and the fluid, as Winternitz and Smith have shown, was rapidly absorbed.

These data it is believed, support the conception of pulmonary edema in the gassed animal, stated in the earlier paragraphs of this discussion, namely, that it is a purposeful reaction and is not directly responsible for the animal's death.


Several investigators, notably Schaefer, Hill, Gunn, Barbour and Williams, have studied the action of chlorine on the bronchi. Their methods have not been uniform, differing not only in the concentration of gas used, but also in the technique of exposure. These differences in method may account for some of the differences in their conclusions. Shaefer, 25 working first with the isolated animal lung and using a chlorine concentration of 1 to 10 or 1 to 20, observed no effect on the bronchial caliber, but in later experiments, with the lung in situ and respiration maintained with the Brodie pump, he noted a definite dilatation of the bronchi. Hill, 6 on the other hand, using weaker concentrations of gas, saw no evidence of bronchial dilatation, but on the contrary thought that there might be some constriction of the bronchioles. In more thorough post-war studies on this problem, Gunn 26 reported as follows:

Generally, the results of experiments with chlorine have gone to show that inhalation of 1 in 5,000 up to 1 in 1,000 produces an increased rate of respiration with a transient broncho-constriction. This broncho-constriction is produced reflexly by the first contact of the irritant vapor with the bronchial mucous membrane. It lasts such a short time that the therapeutic measures to combat it would be unavailing. A subsequent sudden increase in the concentration of the gas may produce another transient reflex broncho-constriction. Continued inhalation of the gas produces thereafter an apparent slight and gradual broncho-constriction, but from histological observations, it appears probable that this is due rather to edema of the bronchial mucous membrane than to contraction of the bronchial muscle.

It is perhaps well to keep in mind that reflex effects arising from irritation of the upper air passages are excluded in these experiments. It is theoretically probable that irritation of the nasal and laryngeal mucous membranes will produce more immediate and possibly more intense reflex effects on the bronchi than will irritation of the trachea, for the sentinel posted in the latter region challenges the enemy gas too late.

A report of the physiology (war) committee of the Royal Society 27 on the pathology and treatment of pulmonary irritant gases contains the following paragraph on bronchial spasm:

In cases of accidental gassing by phosgene or chloropicrin in factories, a frequent an distressing symptom is spasmodic asthma, which may recur even after brief inhalation of small


quantities of vapor. Well-marked constriction of the bronchioles has been observed as the result of inhalation of phosgene in animals (Golla). Exposure to chlorine has not been observed experimentally to produce this effect (Bayliss). It is not yet clear how far and for how long bronchial spasm may recur in men gassed in warfare. Laryngeal spasm may sometimes have been confused with bronchial spasm in man.

Barbour and Williams, 28 working in the Yale laboratories, exposed freshly excised bronchi of dogs and calves to the action of chlorine in Locke's solution, and noted that while low dilutions of the gas produced in some cases a transient relaxation of the bronchial musculature, concentrations of 200 milligrams to the liter, or more, regularly produced well-marked constriction which might or might not be preceded by a transient relaxation. These observers also noted similar reactions in the smooth muscle of the veins and arteries, but the constriction-producing concentration was found to be considerably lower in the case of the bronchi than in the vessels.

Both grossly and microscopically, there was seen, at autopsy in chlorine gassed dogs dying acutely, a dilatation of the bronchi of varying degree. But the dilatation was not pronounced and might have been due to mechanical obstruction of the bronchi by sloughs of necrotic membrane and exudate rather than to a direct action of the gas on the bronchial wall. In the delayed deaths, and more particularly in dogs recovering from severe gassing and killed some time afterwards, bronchial dilatation, or rather true bronchiectasis, constituted a very conspicuous feature of the lungs. In these cases, it seemed quite clear that the chronic infection with weakening of the bronchial wall, aided by obstruction resulting from the organizing bronchiolitis, was the chief etiologic factor.

Microscopically, changes suggesting a nodal constriction of the bronchioles were occasionally noted in animals dying acutely from both chlorine and phosgene, but upon careful study it was not determined that such pictures might not have been produced by twists or angles in the bronchial tree. Careful reconstructions would be necessary to settle this point.


In discussing the subject of circulatory disturbances following gassing several questions, which are more or less closely interrelated, must be considered. Do the respiratory irritant gases directly damage the wall or contents of the pulmonary capillaries? If so, what is the nature of the injury or disturbance and what anatomical evidence is there that such changes affect the general pulmonary and systemic circulations?

Since the answer to the first question, to which the others are subsidiary, must be a qualified one, it is clear that much of the discussion of this problem must be to a certain extent speculative. There are, however, certain observations and deductions upon which most investigators are agreed.

For example, it is well established that exposure to very high concentrations of chlorine, phosgene, or chloropicrin may cause death in a few minutes, and that dilatation of the pulmonary vessels with stasis is the chief finding in such cases. Whether death under these conditions is due to a diffuse injury to the respiratory epithelium, or to vascular injury with dilatation, stasis, and consequent asphyxia, can not be stated, but the latter explanation appears the more plausible.


It is, however, in cases in which marked symptoms and death come on after some hours that the problem of circulatory disturbance is of most interest, and it is here that one finds disagreement as to the fundamental facts, particularly regarding the changes in the pulmonary vessels.


Klotz, in studying the action of chlorine on rabbits and mice, observed hyaline thrombi in the pulmonary capillaries and regarded them as the cause of serious obstruction of the blood flow through the lungs: 1

An important observation in the acute deaths, particularly in mice, was the finding of patches of diffuse coagulation of blood in the pulmonary capillaries. In these areas, we have observed within the capillaries and larger vessels the presence of a diffuse meshwork-like altered fibrin. Wide stretches of channels were found in which an irregular meshwork of threads stain diffusely blue with hematoxylin. In these thrombi relatively few red blood cells were found. Similar coagula with a varying number of erythrocytes were seen in the arterioles and venules. This process of thrombosis was not uniformly distributed through the lung.

Klotz 1 believes that embarrassment to the organism results from both failure of the right side of the heart and deficient oxygenation. Dunn 9 found in goats, killed by phosgene, capillary thrombi similar to those observed by Klotz in rabbits. In describing the changes in the initial period, that is, the first few hours after gassing, Dunn stated:

The cardinal feature is damage of capillary blood vessels in the zone where these are first exposed to the gas, with no greater protection than is afforded by the delicate pulmonary epithelium. The most striking evidence of injury of capillaries is the formation of thrombi in their lumina. When thrombosis is fully developed, it entails blockage of the affected vessels, so that the progressive exudation of fluid which continues long afterwards, must presumably occur from other vessels in which no microscopic change can be recognized.

In a discussion of the later changes, Dunn added:

The thrombosis of the capillaries undergoes very little extension after the end of the initial period and persists practically unchanged for 36 to 48 hours. It is always accompanied by much engorgement of the neighboring capillaries. * * * In animals living after 48 hours, there may be observed some absorption of the minute thrombi by phagocytes.

Perfusion of the lungs with salt solution failed to dislodge the thrombi, and the subsequent injection of carmine gelatin mass demonstrated that there was in places definite edema of the capillaries. The obstruction, however, was focal, rather than diffuse, large sections of the lung being well injected.

Ricker, 29 on the basis of extensive studies upon the action of phosgene on cats, emphasized the importance of dilatation of the vessels with resulting stasis, but did not attribute the disturbance to capillary thrombosis.

Similarly "Winternitz and his coworkers have been unable to convince themselves that thrombosis occurs as a consequence of the injury by any of the pulmonary irritants, although they point out that the presence of a layer of fibrin along the alveolar walls in many cases, with threads crossing the septa (fig. S7; also see figs. 66 and 67) must constitute a definite barrier to the free flow of blood through the lung. l
l Regarding the presence of thrombi in human lungs, following gassing, Pappenheimer says: "Our preparations show no clear evidence of thrombus formation, either due to agglutination of red blood cells, to ante-mortem fibrin deposition or to the agglomeration of platelets in the pulmonary capillaries."


In view of the difficulty in recognizing a freshly formed capillary thrombus, which has little to distinguish it from a post-mortem clot, and lacks the character structure of mural thrombi in large vessels and heart, these conflicting observations are understandable.

As regards the significance of injection studies, it may be pointed out that, as Kline and Winternitz 30 have shown, great difficulty is found in perfusing and injecting the lungs in human lobar pneumonia, where outpouring of fibrin takes place even greater than that seen in gassed animals. The circulatory disturbance is probably much the same in the two conditions, in other words, equally obscure in etiology.

FIG. 87.- Thick layer of fibrin along the alveolar wall seven days after phosgene gassing


Much emphasis has been placed on the dilatation or at least the distention of the heart at autopsy, in acute deaths from gassing. In a report of the physiology (war) committee of the Royal Society, the following statement is made with reference to this findings: 31

In fatal cases dying within 24 hours, the heart is usually found distended to a considerable degree with fully clotted blood. This is observed even when the post-mortem examination is made within two hours after death. When death occurs later, in about three days, dilatation is usually, though not invariably, observed.

In dogs the distention, of the right ventricle particularly, was often quite striking, and most observers have regarded it as an evidence of cardiac failure. Clinical observations, both in man and animals, have tended to confirm this impression. For example, according to Peters, 31 among gassed soldiers examined within 24 hours after gassing, cardiac enlargement to the right was a


common finding; in some cases the heart border was as much as 1 inch to the right of the sternum. By X-ray photographs Bunting demonstrated a definite increase in cardiac outline in dogs gassed with phosgene. The discussion of the significance of these intravitam observations does not lie within the scope of this chapter. But a recent study of the state of the heart at autopsy in cases where death was apparently not due to cardiac failure, suggests caution in attaching particular significance to post-mortem cardiac distention. Some of these observations may be briefly recorded. Six dogs, dying apparently from malnutrition in the course of feeding experiments, showed no general signs of circulatory disturbance; the heart in all cases, however, was more or less distended, as the measurements in Table 61 show:

TABLE 61.- Dogs dying of malnutrition

The heart measurements m in six dogs of similar size, which died within 24 hours after chloropicrin gassing, are shown in Table 62:

TABLE 62.- Dogs dying of chloropicrin poisoning

A comparison of the figures in the two series shows a slightly greater average cardiac capacity in the nongassed dogs. It might be added that a review of the findings in 20 additional chloropicrin dogs showed that the cases cited were representative. The average cardiac capacity of the entire series was, in fact, somewhat lower than that of the group from which figures are given.

In some observations on normal dogs killed by chloroform, the heart showed practically the same degree of distention as in the fatally gassed animals.

These findings do not, of course, invalidate clinical observations regarding the state of the heart during life, either in gassed animals or man, but they do suggest caution in the interpretation of autopsy findings with respect to cardiac dilatation.

m Upon exposing the heart the great vessels were carefully clamped, then tied off with heavy cord and cut fairly close to the heart. Practically no blood was lost if care was exercised, and the difference in the weights of the full and empty organ showed very accurately the cardiac capacity.



Reference has already been made to hemorrhages in the pleura and endocardium of animals dying acutely from exposure to chlorine, phosgene, chloropicrin, and other gases of the respiratory irritant group. n In respect to size, location, and distribution these hemorrhages did not differ essentially from those seen in man in association with some of the severe infections or toxemias. In the lungs they were often overlooked at autopsy by the casual observer, owing to the extreme congestion and edema present, which obscured the picture. The posterior surface of the lung was the site of predilection, but gross extravasations were seen occasionally in the substance of the organ. Small foci in the lung, overlooked grossly, were frequently discovered upon microscopic examination.

The subendocardial hemorrhages were very conspicuous and not easily overlooked. They occurred most often in the left ventricle, were generally multiple, and distributed lengthwise along the crests of the muscular pillars and ridges. Now and then they were found about the bases of the valve cusps, mitral, tricuspid, and semilunar.

The hemorrhages in these acute deaths, though very striking in many cases, are of less significance than those seen in the chronic or recovered animals. It is the tendency to hemorrhage in this second type of animals that it is proposed to discuss here in some detail.

In the first series of autopsies on animals which had died five days to several months after gassing with phosgene and chlorine, hemorrhages in the lungs were recorded in approximately 35 percent. The cause of death in a majority of these cases was respiratory infection of one type or another, generally bronchopneumonia, associated with an acute and chronic bronchitis.

In a second series of autopsies on "chronic dogs" that were killed 10 days to 5 months after gassing, for the purpose of studying the residual changes in the respiratory system, it was noticed that hemorrhages in the lungs were much more regularly encountered, the percentage being 90 as compared with 35 in the animals that died. Furthermore the hemorrhages were larger and more widely distributed. In most cases they were quite fresh, indicating that they had occurred just before death. These dogs had been killed with strychnine and all died in convulsions, which sometimes lasted several minutes.

In order to determine to what extent the convulsions were responsible for the hemorrhages, other methods of killing were resorted to. The dogs were divided into three groups. The first were killed with potassium cyanide (subcutaneous injection); the second with chloroform (forced inhalation); the third by a shot through the head from a small caliber pistol.

The dogs killed with chloroform were forced to breathe through an ordinary anesthetizing cone. Most of the animals struggled considerably. Hemorrhages in the lungs were found grossly in 60 per cent, and this figure was increased to 65 percent by a microscopic study of the lungs.

The cyanide dogs died within a few minutes after the injections. Some showed only a slight rigor; in others there were definite convulsions, but never as marked or prolonged as in the dogs killed with strychnine. At autopsy the lungs showed hemorrhages in approximately 50 percent.
n Punctiform hemorrhages in the brain in fatal eases of phosgene gassing in man have been described by Mott (Brit. Med. Jour., 1917, i, 637) and by Ricker (Samml. Klin. Vortr. Inner. Med., 1919, Nos. 256-260, p. 727). Similar lesions have not been recorded in gassed animals, possibly because the brains in most cases have not been carefully examined.


The dogs killed by shooting succumbed with less struggle than those killed with strychnine, chloroform, or cyanide. Hemorrhages were demon- strable in only 30 percent, and in some of the positive cases the lesions were obviously not terminal.

From these observations, it may be concluded: (1) That hemorrhages occurred in a little less than 50 percent of the "chronic" gassed dogs which died from respiratory infection or other cause; (2) that hemorrhages were present in a still higher percentage in dogs that were killed, the percent varying with the method of killing, strychnine giving almost 100 percent, shooting 30 percent. In brief it appears that the lung of an animal that has been gassed is a favorable site for hemorrhage, and that hemorrhages into such a lung are easily induced by struggling or convulsions.

It should be added that 15 normal control animals were killed by the methods mentioned, and that while strychnia and chloroform produced hemorrhages in some cases, the lesions were never as extensive as in the gassed dogs.

The character of the pulmonary hemorrhages in these gassed dogs and the relation to the chronic inflammatory or reparative lesions so regularly found in the lungs of recovered animals are points which have interested us particularly. A brief description of the hemorrhages will suffice.

Grossly, the hemorrhages often resembled those seen in the acute stage; that is, they appeared as irregular subpleural extravasations of variable size and number situated generally on the posterior surface of the lung. But more often they occurred in the substance of the lung as nodules, generally irregularly spherical in outline. They varied in diameter from a millimeter to several centimeters. The larger ones were surrounded by a light zone of noncollapsed lung tissue, evidently the result of pressure of the hemorrhage on a bronchus. The relation of some of the smaller hemorrhages to bronchi was often quite clear, particularly in dogs killed in the subacute period, 5 to 10 days after gassing. The hemorrhagic foci in these cases were seen to coincide with the fine nodules of organizing bronchiolitis. (See PI. XIII.)

Now and then an extravasation of blood was seen forming a sort of halo about a fair-sized artery in the lung. (Fig. 88.) This is the so-called "ring hemorrhage" which was obviously the result of the rupture of one of the vasa vasorum, with the escape of blood into the perivascular sheath. Similar hemorrhages have been described by Mott 7 in the human brain, following fatal phosgene gassing, and by Ricker 29 in cats experimentally gassed.

The question naturally arises as to whether there is any relationship between these hemorrhages and the chronic infection and reparative changes which characterize the gassed lung. In answer it can be stated, on the basis of microscopic examination of a large number of lungs, that in practically every case in which hemorrhages were found, residual focal lesions, generally of the nature of an organized bronchiolitis, were demonstrable. Furthermore, in many cases, especially in dogs killed in the subacute period (5 to 10 days after gassing), the hemorrhages were seen to originate directly in these lesions. (See PI. XIII.)

In addition to fresh hemorrhages one saw not infrequently patches of disintegrating blood cells and scarred areas containing blood pigment. The presence of these would suggest that hemorrhages were not only the effect of


a scarred damaged lung, but might also have been the cause. The delicate vessels of the scar were prone to rupture. The resulting hematoma became organized with the formation of more scar tissue. In other words, it seems possible that in these chronic lesions of the lungs and in hemorrhages there may have been a vicious circle, with sudden increase in blood pressure (induced by excessive physical effort, struggling convulsions, or other cause), as the force which kept the process going.

The conclusions to be drawn from the foregoing observations upon hemorrhages in the lungs of chronic gassed animals are: (1) The chronic inflammatory and reparative changes in these lungs created a condition favorable to hemorrhage; (2) hemorrhages were induced in such lungs by sudden increase in blood

FIG. 88.- Hemorrhage into a perivascular sheath of a pulmonary vessel in a dog dying 10 days after phosgene gassing

pressure; (3) hemorrhages might lead to further scarring of the lung, which again favored bleeding, thus constituting a vicious circle.


A marked tendency to respiratory infection, after exposure to irritant gases, has been observed in both man and animals. The most outspoken examples of such infection were seen in cases where the gas injury was severe, but not sufficient to cause death within 24 hours. A widespread necrotizing bronchopneumonia was the usual result in these cases.

Several questions present themselves in respect to the origin of the infection and its manner of spread.

Even within a few hours after gassing, particularly in the case of chlorine, a definite inflammatory reaction was seen in the lungs. Apart from the edema, the inflammatory nature of which may be considered debatable, one found in


FIG. 89.- Necrosis of bronchial epithelium with acute inflammatory reactiono24 hours after chlorine gassing


the alveoli, small quantities of fibrin, few scattered polynuclear leucocytes, and occasional large mononuclear (epithelial) cells, and in the capillaries a definite polynuclear cell increase. This mild reaction was fairly diffuse and may be interpreted as a reaction to the chemical injury.

Six to twelve hours after gassing (often longer) another type of reaction, clearly focal in origin, was frequently seen. The chief element in the exudate was the polynuclear leucocyte, and the location was most often in and about the bronchiole. (See figs. 54 and 68.) In the case of chlorine a similar reaction was found also in the walls of the large bronchi and trachea. (See fig. 61.) The question has been raised as to whether this early inflammatory process was a response to the chemical injury or to a complicating infection. In view of the difficulty in demonstrating bacteria in sections at this stage, it has been argued that the reaction was a direct response to the gas injury. It seems to us, how- ever, that for several reasons, this interpretation of the reaction can not be maintained: 1. The difficulty of demonstrating bacteria in sections is well

FIG. 90.- Wall of small bronchus showing mucosa entirely destroyed by phosgene and a large bronchus of same animal with uninjured mucosa

recognized, and negative findings are therefore not conclusive. 2. Cultures show that bacteria are present in the lungs of a high percentage of gassed animals, and in practically all cases in which a frank pneumonic reaction has developed. 3. The character of the exudate, which is predominantly polynuclear, the focal distribution of the lesions, and the localization about the atria suggest a reaction to a bacterial rather than to a chemical injury.

Furthermore, it has been pointed out (Smith) that the similarity of the organisms found in the lungs of gassed animals with those present in the mouths of normal dogs indicates that the organisms make their way from the mouth to the lung shortly after gassing. In support of this idea the following observations on two series of chlorine gassed dogs are presented: 32

In the first series, the lungs from 25 animals were cultured. These include dogs that lived 12 hours to 48 days after exposure. In 6 cases, cultures were negative. In 1, a pure culture of a small Gram-negative hemoglobinophilic bacillus was recovered. In 12, pure cultures of pneumococci resulted, and 6 cases, both the pneumococcus aln the small Gram-negative bacillus were


obtained. In five of the animals post-mortem blood cultures were also made. Two of the animals had died 24 hours after gassing, one after 3 days, one after 6, and one after 36 days. In all of these, pure cultures of a Gram-positive diplococcus, similar to that found in the lungs, were recovered.

A more detailed study was subsequently carried out on a series of 21 dogs. In these the normal flora of the mouth was first determined. The cultures were made on two successive days, prior to gassing. In addition to such common organisms as streptococci, staphylococci, and B. subtilis, there were found in the mouth of each of the dogs, a Gram-positive diplococcus that agglutinated with pneumococcus group 2 serum, at a dilution of 1 to 2, and a small Gram- negative hemoglobinophilic bacillus very similar to the influenza bacillus. In

FIG. 91.- A comparison of the injury to the tracheal mucosa by chloropicrin, phosgene, and chlorine. A, Chloropicrin, with damaged hut intact epithelium; B, chlroropicrin with sloughing of superficial epithelial layer; C, phosgene, with undamaged mucosa; D, chlorine, with killed and exfoliated mucosa

14 cases both of these latter organisms were found, while in 6 the pneumococcus alone, and in the Gram-negative bacillus alone was obtained.

The lung cultures taken post-mortem were negative in 4 cases where death resulted shortly after exposure. In 9 cases the pneumococcus alone was obtained. In 4 the pneumococcus and the Gram-negative bacillus were recovered, and in 4 others the Gram-negative bacillus alone was recovered.

Blood cultures (post-mortem) were positive in 5 of the 21 animals. The organism recovered in each case was the pneumococcus.

The conclusion from this study is that organisms which normally inhabit the mouth of the dog find their way into the lungs shortly after gassing and remain there for a long time in animals that survive the acute stage.



Animals recovering from moderately severe exposures to any of the pulmonary irritant gases showed, even up to five months after gassing, definite residual lesions in the lungs of greater or lesser degree. o These changes have been fully described and illustrated in the reports on the pathology of chlorine and phosgene gassing. They probably occurred quite as regularly after exposure to other gases, which have not been as fully studied. The question to be considered here is that of the significance of the lesions, and particularly their bearing on certain clinical phenomena, effort-syndrome, tendency to chronic respiratory infection, etc., observed in human gassed cases and regarded as late effects of the gas injury.

Subjective symptoms in dogs obviously did not admit of investigation, but the general appearance of the animals, weakness, loss of weight, tendency to cough on exertion, suggested a condition in the recovered gassed animal similar to that in man. No reports have been found of systematic clinical or physiological studies of animals such as have been made by Lewis, Barcroft, Pearce, and others in man.

Regarding the nature of the anatomical findings in these dogs, little need be added to what has gone before. They varied in individual animals, and to some extent with different gases. In the more pronounced cases, which ended fatally, there was a suppurative bronchitis and bronchiectasis, with occasionally a chronic infection of the lung tissue itself, that is, an organizing or interstitial pneumonia. In the majority of animals, however, the lesions were healing or healed, and were limited to certain of the smaller air passages and the lung tissue supplied by these. The bronchioles were partially or completely obliterated; the tributary lung tissue was atelectatic or emphysematous.

The frequency and general character of the changes found in recovered dogs are very well shown in Table 50 (p. 429) for chlorine, and Table 54 (p. 450) for phosgene. It is seen that of the chlorine dogs which died or were killed from 15 to 193 days after gassing, 51 percent showed bronchitis of one type or another, and 24 percent pneumonia. Organization was a feature of both the pneumonic and bronchial lesions. Of the phosgene dogs, 60 died and 103 were killed two weeks to four months after gassing. Bronchitis was demonstrated in 68 percent, pneumonia in approximately 10 percent. These figures are based on gross findings confirmed by one or more microscopic sections. It is very certain that a more thorough microscopic examination in all cases would have increased these percentages considerably. Indeed, after reviewing a large amount of material it is our impression that a severe exposure to practically any of the respiratory irritant gases--phosgene, chlorine, and chloropicrin particularly--makes for a permanent and usually demonstrable damage to some part of the respiratory tract. The effects on the respiratory tract of repeated (sublethal) exposures to irritating gases has not been systematically investigated. But it has been the impression, from a routine examination of the organs of a number of dogs killed after two or three exposures to chlorine or phosgene (intervals of ten days or more), that the residual or chronic changes
o The organic arsenicals would appear to be exceptions to this general statement in that according to Finney's studies none of the recovered dogs showed any residual lesions. In view of the fact, however, that such lesions were found after mustard gassing, the effects of which are so similar to those of the arsenicals studied, one is inclined to think that a more thorough investigation of Finney's cases wotild have shown chronic pulmonary changes.


in the lungs were much more marked than in dogs recovering from a single gassing. There is, of course, nothing remarkable in this observation; although it is of interest here that Underhill 5 did not observe, in the case of phosgene at least, that animals previously gassed were more susceptible to a second gassing, as judged by sumptoms and mortality figures.

If lesions analogous to those described in dogs follow gassing in human beings (and there is no reason to suppose that they do not) there would appear to be a clear anatomic basis for the clinical phenomena observed. In this connection the following quotation regarding changes characteristic of the period of convalescence, from a report compiled by the physiology committee of the Royal Society, 28 is of interest.

Clinical observations (T. Lewis, Riddell, Price, Jones, and Hunt) indicate that, temporarily, as large a proportion as 36 percent of their gassed subjects may exhibit the so-called "effort- syndrome." There seems no sufficient reason to assume in such subjects a specific change in the tissues of the lungs which might be held accountable for such a symptom as breathlessness; in some cases, though not in all, there are grounds for believing that in the earlier asphyxial stage the heart may become permanently strained. There is no evidence of a permanent dilatation of the organ, though this point has been carefully investigated. In the various cases admitted into the special hospitals for the heart, gassing is regarded as a relatively unimportant cause, and where it occurs in the histories it is stated that in as many as 30 percent the men were at the time of gassing probably affected with the condition which exhibits the "effort-syndrome " (T. Lewis). The condition, as far as gassing is concerned, must be regarded as a feature of convalescence, and not as a persistent stage.

The question of cardiac dilatation after gassing has been discussed in another paragraph; here the doubtful value of post-mortem evidence is pointed out. A review of the figures upon the weights of the unopened and empty hearts in a large series of gassed dogs, shows that, in general, the capacity of the heart is greater in delayed than in acute deaths, but, as previously emphasized, the figures are practically the same as for animals dying from other conditions, so that it seems unjustifiable to draw any conclusions from these data.

The pulmonary lesions in dogs were clearly regressive, just as the tendency to recovery indicates that they are in man. Ten days after severe but not fatal gassing with phosgene or chloropicrin, the dog's lung was studded with readily palpable and visible tubercle-like nodules of organizing bronchiolitis, whereas a month after gassing these foci were practically microscopic, unless, as happened in a certain percentage of cases, a complicating infection had persisted.

Definite evidence of the persistence of pathogenic microorganisms in the lungs of these recovered animals was obtained by making lung cultures in 31 cases. The dogs had recovered from a moderately severe exposure to phosgene and were killed at varying intervals after gassing. Cultures from the small bronchi and lung tissue itself were taken promptly after death, using broth and blood agar. The findings were as follows:

Staphylococcus aureus.................................................5
Streptococcus hemolyticus...........................................5
Streptococcus nonhemolyticus......................................2
B. subtilis.....................................................................4
Miscellaneous organisms..............................................5
No growth.................................................................10


It is seen that the lungs were sterile in less than a third of the cases, while pathogenic organisms were present in approximately 50 percent.

From these observations it may be inferred, therefore, that the lung of the gassed animal or human being is the seat of chronic infection which may at any time become active and dangerous. p It is clear also that the disability manifested by a certain number of gassed individuals may be due partly to mechanical interference with proper respiratory exchange from bronchiolar obliteration, emphysema, etc., and partly to the noxious effects of a persistent respiratory infection.

p Since our observations were made chiefly on dogs, they throw no light on the recently much debated question as to the effect of exposure to respiratory irritant gases on the development of pulmonary tuberculosis. It may be pointed out that the more recent clinical studies on latent effects of gassing in man, notably those of Achard, fail to show any etiologic relation between gas injuries and tuberculosis


(1) Klotz, Oskar: Acute Death from Chlorine Poisoning. Journal of Laboratory and Clinical Medicine, St. Louis, Mo., 1917, ii, No. 12, 889.
(2) Edkins, J. S., and Tweedy, N.: The Early Changes Produced in the Lungs by Gas Poisoning and Their Significance. Reports of the Chemical Warfare Medical Committee (British), April, 1918, No. 2, 4. On file, Technical Division, Chemical Warfare Service.
(3) Wagner, J. H.: Bronchiolitis Obliterans Following the Inhalation of Acrid Fumes. The American Journal of the Medical Sciences, Philadelphia, Pa., 1917, cliv. No. 4, 511.
(4) Hoover, C. F.: Phosgene Poisoning (unpublished report).
(5) Underhill, F. P.: The Physiology and Experimental Treatment of Poisoning with Lethal War Gases. Archives of Internal Medicine, Chicago, Ill., 1919, xxiii, No. 6, 753.
(6) Hill, L.: Gas Poisoning: Physiological Symptoms and Clinical Treatment. Lancet, London, 1915, ii, December 4, 1248.
(7) Mott, F. W.: Punctiform Hemorrhages of the Brain in Gas Poisoning. British Medical Journal, London, 1917, i, May, 637.
(8) Winternitz, M. C., Finney, W. P., jr., and Wislocki, G. B.: Note on the Comparative Pathology of Acute Phosgene Poisoning. Collected Studies on the Pathology of War Gas Poisoning. Yale University Press, New Haven, Conn., 1920, 63.
(9) Dunn, J. S.: Report on Pathological Changes in the Lungs of Goats Resulting from Inhalation of Phosgene. Reports of the Chemical Warfare Medical Committee (British), September, 1918, No. 9, 4. On file, Technical Division, Chemical Warfare Service.
(10) Winternitz, M. C., and Wislocki, G. B.: The Pathology of Superpalite Poisoning. Collected Studies on the Pathology of War Gas Poisoning, Yale University Press, New Haven, Conn., 1920, 89.
(11) Lambert, R. A., and Jackson, L.: The Pathology of Chloropicrin Poisoning. Collected Studies on the Pathology of War Gas Poisoning, Yale University Press, New Haven, Conn., 1920, 69.
(12) Winternitz, M. C., and Smith, G. H.: Preliminary Studies in Intratracheal Therapy. Collected Studies on the Pathology of War Gas Poisoning, Yale University Press, New Haven, Conn., 1920, 143.
(13) Winternitz, M. C., Finney, W. P., jr., and Wislocki, G. B.: The Pathology of Cyanogen Chloride and Bromide. Collected Studies on the Pathology of War Gas Poisoning, Yale University Press, New Haven, Conn., 1920, 115.
(14) Geppert, J.: Ueber das Wesen der Blausaurevergiftung. Zeitschrift für Klinische Medizin, Berlin, 1888-89, xv, 307.
(15) Kobert, Rudolf: Lehrbuch der Intoxikationen, Ferdinand Enke, Stuttgart, Germany, 1893, 513. Textbook.
(16) Marshall, E. K., jr., and Miller, E. J.: Chemical Warfare Service Reports. On file, Technical Division, Chemical Warfare Service.
(17) Wislocki, G. B.: The Pathology of Poisoning by Arsine. Collected Studies on the Pathology of War Gas Poisoning. Yale University Press, New Haven, Conn., 1920, 129.
(18) Mackenzie, G. M.: The Action of Arsine on the Blood, 1918. (Report not published.)
(19) Clark and Pappenheimer: (Report not published.)
(20) Dunn, J. S.: Report on the Nature of Lesions Produced in Experiments on Animals by Inhalation of Diphenylchlorarsine and Allied Compounds. Reports of the Chemical Warfare Medical Committee (British) September, 1918, No. 9, 35. On file, Technical Division, Chemical Warfare Service.
(21) Finney, W. P., jr.: The Pathology of Poisoning by Organic Arsenic Compounds. Collected Studies on the Pathology of War Gas Poisoning, Yale University Press, New Haven, Conn., 1920, 135.
(22) Pearce, L., and Brown, W. H.: Chemopathological Studies with Compounds of Arsenic. I. Types of the Arsenic Kidney. Journal of Experimental Medicine, New York, November 1, 1915, xxii, No. 5, 517.
(23) Barcroft, J.: Report on the Condition of the Blood in Cases of Gas Poisoning. Report of the Chemical Warfare Medical Committee (British) April, 1918, No. 6, 4. On file, Technical Division, Chemical Warfare Service.
(24) Eyster, J. A. E.: Report of the Chemical Warfare Medical Committee, Medical Research Commission (unpublished).
(25) Schaffer, E.: On the Immediate Effects of the Inhalation of Chlorine Gas. British Medical Journal, London. 1915, ii, August 14, 245.
(26) Gunn, J. A.: The Action of Chlorine on the Bronchi. Report to the Chemical Warfare Committee. Quarterly Journal of Medicine, Oxford, England, 1920, xiii, January 19, No. 50, 121.
(27) Report of War Physiology (War) Committee of the Royal Society, September, 1917, and adopted by the Chemical Warfare Medical Committee (British) January, 1918, No. 1, 7. On file, Technical Division, Chemical Warfare Service.
(28) Barbour, H. G., and Williams, H. W.: The Effects of Chlorine upon Isolated Bronchi and Pulmonary Vessels. Journal of Pharmacology and Experimental Therapeutics, Baltimore, Md., 1919, xiv, September, No. 1, 47.
(29) Ricker, G.: Beiträge zur Kenntnis der Toxischen Wirkung des Chlorkohlenoxydgases (Phosgene). Sammlung Klinischer Vorträge Innere Medizin, Leipzig, 1919, Nos. 256-260, 727. (30) Kline, B. S., and Winternitz, M. C.: Studies upon Experimental Pneumonia in Rabbits. VIII. Intra Vitam Staining in Experimental Pneumonia, and the Circulation in the Pneumonic Lung. The Journal of Experimental Medicine, New York, 1915, xxi, No. 4, 311.
(31) Peters, R. A.: Changes Observed in the Heart and Circulation as a Result of Gas Poisoning. Report of the Chemical Warfare Medical Committee (British) April, 1918, No. 7, 5. On file, Technical Division, Chemical Warfare Service.
(32) Winternitz, M. C., Lambert, R. A., Jackson, L., and Smith, C. H.: The Pathology of Chlorine Poisoning. Collected Studies on the Pathology of War Gas Poisoning. Yale University Press New Haven, Conn., 1920, 28