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 X




Among the problems submitted to the medical division of the Chemical Warfare Service was the study of the changes induced by toxic gases in the economy of animals, with a special reference to modifications in intermediary metabolism, and the determination of alleviative and curative measures.

It is obvious that in order to interpret intelligently the effects of gas poisoning upon the organism, it is essential to determine the changes, so far as possible, that the gases bring about. Inasmuch as these alterations are generally of a chemical nature, the problem must be attacked by chemical methods. The knowledge thus gained is fundamental for any measures that may be em- ployed to prevent or alleviate the detrimental effects of poisonous gases.

In order that consistent results might be obtained, considerable attention was given to the experimental method of gassing. The technique as finally evolved was worked out in cooperation with officers of the Chemical Warfare Service. In principle it consisted of passing a mixture of gas and air through an air-tight chamber containing the " experimental animal" for a definite period at a determined rate, checking the mixture by frequent analysis of samples taken from the chamber. Unless otherwise indicated, the dog was employed as the experimental animal throughout the investigation.

The exposure of the dog to the gases elicited the following reactions, each gas differing slightly from its neighbor. The general clinical symptoms included by gassing with chlorine were, at first, general excitement, as indicated by restlessness, barking, urination, and defecation. Irritation was distinctly visible, as indicated by the blinking of the eyes, sneezing, copious salivation, retching, and vomiting. Later the animal showed labored respiration, with frothing at the mouth. Food was refused, although the animal might drink a large quantity of water. The respiratory distress increased until eventually death occurred from apparent asphyxiation. On the other hand, if the concentration of gas was not lethal the animal would present an emaciated appearance and be greatly distressed for several days, followed by ultimate recovery and return to apparently normal conditions.

Phosgene acted chiefly as a respiratory irritant, but was also a lacrymator. Very small doses, scattered in the air, caused coughing, watering of the eyes, and intense dyspnea. It differed from chlorine in that in these small concentrations its influence was limited mainly to the terminal air cells of the lungs. This effect led to edema of the lungs, accompanied by interference with the respiratory exchange and consequent cyanosis, a grave condition usually terminating in death. The first symptoms were dizziness and cyanosis on exertion. It usually required several hours for the serious symptoms to develop, and in the interval there might have been no sign of danger.

a The data in this chapter are based, in the main, on the experimental observations made by the section on intermediary metabolism of the medical division of the Chemical Warfare Service at Yale University, account of which is found in "The Lethal War Gases," by Frank P. Underhill, New Haven, Yale University Press, 1920.


At high concentrations there was slight lacrymation and uneasiness. The pupils became clouded, but the animal showed no violent symptoms. Subsequently dogs exposed to high concentrations developed a hard cough, respiration became more and more difficult, usually there was a rattling in the throat, and death followed three to twelve hours after exposure.

Animals subjected to lower concentrations developed the same chain of symptoms which, however, were not quite so severe at first. Death occurred, as a. rule, after 18 to 36 hours. After death the nostrils and trachea were filled with mucus. The lungs were collapsed and filled with mucus and blood. The slow filling of trachea and lungs accounted for the deaths which occurred 18 to 48 hours after gassing. The heart action grew weaker as death approached but persisted after all attempts at breathing had ceased.

Chloropicrin is a lacrymatory and respiratory irritant. Repeated exposure was said to cause increased susceptibility. Exposure to this gas produced coughing, nausea, and vomiting, and in large quantity could cause unconscious- ness. Secondary effects were bronchitis, shortness of breath, a weak irregular heart, and gastritis. Chloropicrin could also cause acute nephritis. Liquid chloropicrin had a corrosive action on the skin, and scratches and abrasions if exposed to chloropicrin fumes invariably became septic and abscess formation sometimes resulted.

During the early part of exposure to chloropicrin the eyes were irritated, and lacrymation occurred. The mucous membrane of the nose and mouth was irritated almost instantly, the animal licking its nose and swallowing frequently. There was always increased nasal secretion and usually salivation, and in cases where salivation was not observed the animal was usually swallowing the saliva. As a rule retching and vomiting occurred 10 to 15 minutes after the beginning of the exposure with the higher concentrations. With lower concentrations the animal did not always retch. Toward the end of the exposure the animal was usually depressed, and in some cases marked paleness of the mucous membrane in the mouth was observed. The respiration was frequently affected early, being somewhat rapid in the early part of the exposure and becoming slower at the end. A tracheal rattle soon developed and respiration was labored and painful. The animal often had a bad cough, and was generally depressed. Convulsions were observed just before death. In case of survival, the dog had symptoms of bronchitis and rhinitis for a few days and then was apparently normal.

A comparison of the three gases showed quite plainly that chlorine had a very strong irritating action, the animal under observation becoming excited and in evident distress. With chloropicrin the character of the reactions produced were very similar to those of chlorine, except in being less pronounced. Phosgene, on the other hand, appeared to cause the animal no immediate distress. Instead of becoming unduly excited the dog lay quietly in the chamber and even when symptoms of poisoning appeared hyperexcitability was not present. It seemed that a certain degree of peripheral anesthesia was present, handling the animal failing to act as a stimulus to muscular activity and to cause the struggling so characteristic with chlorine and chloropicrin dogs.



The method of exposing experimental animals to definite concentrations of the gas has been indicated. The animals were always gassed singly, experience having demonstrated that when two dogs were gassed at once in the same chamber very inconstant toxicity figures were obtained. This was due to the fact that the gas mixture in the chamber showed very wide variations in composition, probably owing to the difficulty of properly controlling the flowmeter. It is possible that an extensive experience with the flowmeters might have obviated this factor. On the other hand, when the dogs were gassed singly the toxicity figures obtained were strikingly constant.

The question of the lethal concentration of the different gases for dogs was next investigated. For this purpose animals of both sexes, various breeds, ages, and states of nutrition were employed, but in every instance the dogs were considered good subjects for experimentation, none having previously undergone any experimental treatment. In all instances the animals were subjected to the action of the gas for a period of one-half hour. the rate of flow of air through the chamber being 250 liters per minute.


For the investigation with chlorine 112 animals were employed. The results of the study are summarized in Table 17, which shows the toxicity of chlorine gas for a one-half hour exposure at various ranges of concentrations, expressed both in milligrams of chlorine per liter of air and in parts of chlorine per million parts of air. The data from this table demonstrate that dogs gassed. with high concentrations (2.53 mgm., or above, per liter) of chlorine gas usually died from the acute effects within the first 72 hours. The small percentage of animals which survived this acute stage usually recovered within a week. The dog gradually developed an appetite and appeared normal, with the exception of some emaciation and laryngitis or bronchitis, either of which could persist for some time.

TABLE 17.- Toxicity of chlorine gas

At lower concentrations (1.90 to 2.53 mgm. per liter) the percentage of recoveries increased rapidly. Another condition stood out prominently at these concentrations, namely, a group of animals which survived for several days. The symptoms were loss of appetite, extreme depression, weakness, and rapidly developing emaciation. Death after this chronic condition must


be differentiated from the acute deaths, since the former were generally due to secondary factors, usually pneumonia of the purulent type. The animals in the chronic condition did not exhibit the acute symptoms, i. e., labored and distressed breathing, after one or two days. Therefore, the third day was arbitrarily chosen as the extreme limit for acute deaths.

At still lower concentrations (1.58 to 1.90 mgm. per liter) the percentage of recoveries increased markedly. The acute symptoms were much less noticeable and recovery occurred more rapidly than at higher concentrations. Concentrations above 2.53 must be regarded as lethal. Concentrations below 1.90 were rarely fatal under the conditions of these studies.

These facts led to a general classification of gassed animals into three groups: Acute deaths, delayed deaths, and recoveries.

Acute deaths.- Animals which succumbed to the immediate effects of the chlorine gas, namely, deaths directly induced by edema of the lungs. The majority of the deaths from this cause occurred within 24 hours after gassing. but some animals survived for 2 or 3 days. However, these animals formed a rather clearly defined clinical group, and experience has shown that all animals dying within three days could be classified together as "acute deaths."

Delayed deaths.- Animals which survived for more than three days after gassing, but which did not recover. In the majority of cases deaths classed as "delayed" resulted from secondary factors, chiefly bronchopneumonia following the subsidence of the acute pulmonary edema. This group of deaths therefore, could not be ascribed directly to the gassing, and thus fell beyond the limits of this investigation.

Recoveries.- Animals which recovered from the gassing with, in some cases, minor secondary symptoms as bronchitis, laryngitis, slight depression, or emaciation.
The delayed deaths and recoveries, therefore, together comprised animals which successfully resisted the direct effects of the gas, and could be grouped together as "survivals'"; that is, having survived the acute period. With chlorine this group included a relatively large number of dogs, whereas with phosgene and chloropicrin the groups classed as "recoveries" and "survivals." were almost identical.

A study of Table 17 shows that at concentrations below 0.81 mgm. of chlorine gas per liter of air not a single acute death occurred among the dogs of the series. At concentrations between 1.27 and 1.90 mgm. from 6 to 10 percent of the animals died acutely, but none before the second day after gassing. Concentrations between 1.90 and 2.22 and between 2.22 and 2.53 mgm. per liter gave about 50 per cent acute deaths, which were fairly equally divided between the first and second days.

Coming to concentrations above 2.53 mgm., the picture changes abruptly, about 90 percent of the 20 animals in the series dying acutely and of these the great majority of deaths occurred on the first day. A closer analysis of the data from gassing for one-half hour at concentrations above 2.53 mgm. per liter of air shows that between 2.50 and 2.85 mgm. the proportion of acute deaths was 87 percent, which was nearly as great as at concentrations between 2.85 and 6.34 mgm., when it reached 93 percent. In view of this result and the fact that the proportion of acute deaths to recoveries increased rapidly and consistently as the gas concentration was increased up to 2.53 to 2.85 mgm., this


concentration must be considered as representing essentially the minimum lethal toxicity. The same conclusion was reached when the delayed deaths and recoveries were analyzed separately.   

Therefore all the data from the study of over 100 dogs gassed for half an hour at various concentrations between 0.16 and 6.34 mgm. of chlorine gas per liter of air clearly indicated that the minimum lethal toxicity of chlorine gas under the conditions of the experiment was between 2.53 and 2.85 mgm. per liter.


Figures relative to the toxicity of phosgene may be found in Table 18, which is a record of experiments on 327 animals. It will be noted there that the periods of death have been divided into those occurring in one day, two days, and three days, and that these deaths are called the " Total acute deaths." Beyond the period of three days death is called " Delayed death."

The concentrations employed were as follows:
41-50 parts of phosgene per million of air (0.17-0.21 mgm. per liter).
51-60 parts of phosgene per million of air (0.22-0.26 mgm. per liter).
61-70 parts of phosgene per million of air (0.26-0.30 mgm. per liter).
71-80 parts of phosgene per million of air (0.31-0.35 mgm. per liter).
81-90 parts of phosgene per million of air (0.35-0.39 mgm. per liter).
91-100 parts of phosgene per million of air (0.39-0.43 mgm. per liter).
101-110 parts of phosgene per million of air (0.44-0.48 mgm. per liter).
111-125 parts of phosgene per million of air (0.48-0.55 mgm. per liter).

From the figures given it may be seen that, as a general statement, the higher the concentration of phosgene the more acute was death, and that for the most part death occurred within the first 48 hours. If animals survived beyond the three-day period they had a very good chance of complete recovery, the delayed deaths not being especially significant.

TABLE 18.- Toxicity of phosgene gas

A point of considerable interest is the fact that the total recoveries at concentrations between 0.31 to 0.35 and 0.35 to 0.39 mgm. per liter were about the same, although very acute death, especially within a period of 24 hours, was much more marked at the higher concentration than at the lower. From these results it may be concluded that the minimum lethal concentration of phosgene is 0.31 to 0.35 mgm. per liter.b

b The results obtained with the higher concentration, 0.39-0.43 mgm. per liter, are difficult of explanation and apparently are anomalous.



The concentrations of chloropicrin employed were as follows:
49-69 parts of chloropicrin per million of air (0.36-0.50 mgm. per liter
70-89 parts of chloropicrin per million of air (0.51-0.65 mgm. per liter).
91-110 parts of chloropicrin per million of air (0.66-0.80 mgm. per liter).
111-131 parts of chloropicrin per million of air (0.81-0.95 mgm. per liter).
132-151 parts of chloropicrin per million of air (0.96-1.10 mgm per liter).
153-172 parts of chloropicrin per million of air (1.11-1.25 mgm. per liter).

Table 19 gives results of observations on 219 dogs. From this table it may be seen that, in general, the higher the concentration of chloropicrin the more acute was death, particularly the deaths within the first 48 hours. The number of deaths occurring within 24 hours at concentrations of 0.96 to 1.10 and 1.11 to 1.25 mgm. per liter were about equal. The minimum lethal concentration of chloropicrin has been taken as 0.81 to 0.95 per liter, where 43 percent of animals receiving this concentration ultimately recovered.

TABLE 19.-Toxicity of chloropicrin gas

A comparison of the toxicity of the three lethal gases shows that in all acute death was a prominent feature. With chloropicrin very few animals died a so-called "delayed death." With phosgene this feature of delayed death was slightly greater, but not especially prominent. It is quite evident that phosgene was by far the most toxic gas, chlorine being the least poisonous, and chloropicrin standing between. With respect to the acute effects of the lethal gases regarding the similarity of the general effects on dogs, as outlined above, and of these gases on men in the field, a striking illustration is afforded by the following paragraph from a captured German medical pamphlet:

The majority of deaths occur during the first 24 hours and in fact during the first 12 hours, with symptoms of pulmonary edema and failure of the circulation. A diminishing number of cases die on the second and third days with accentuation of the inflammatory symptoms in the lungs. The number of cases that die still later is proportionally very small. A case who has developed no severe symptoms by the third day is seldom endangered. On the other hand, the possibility of a late increase in the gravity of the case can not be excluded with certainty before the end of the first week.




The immediate effect of phosgene poisoning was to cause an increase in the rate of respiration and from a normal figure of approximately 20, the rate during the first few hours after gassing was found in most cases to have risen to about double the normal. In the first few hours following gassing in the animals that were not seriously affected the rate of respiration remained somewhat above normal. In the animals which were seriously affected respiration in general increased in rate but decreased in depth so that there was rapid shallow breathing. This was apparently coupled with the development of edema in the lungs of the animals. The results attained from the study of the respiration apparently did not afford a consistent index as to the condition of the animal, but a rapid shallow breathing, in many of the cases, indicated a serious condition.

The general effect of phosgene gas on the respiratory tract of the animals was not very irritating, in fact, as has been noted by other investigators, the phosgene in many cases appeared to have an anesthetic effect. The animals lay quietly unless they were in very bad condition, very little mucus was given off from the linings of the respiratory tract, and in most cases only slight salivation occurred.

The immediate effect of gassing with chloropicrin was to lower somewhat the normal rate of respiration. Within two to three hours after exposure to the gas the respiration rate regained the normal and in serious cases continued to increase so that within the first few hours a level could be reached which was considerably above the normal. The results from the study of the respiration data as far as attained did not afford a consistent index of the animal's condition.

Chloropicrin was very irritating in its effect upon the respiratory passages and as a result a large amount of mucus and saliva was given off during the first few hours subsequent to gassing. The respiratory passages became more or less clogged and the animal exhibited considerable difficulty in breathing. In many cases the nasal passages were almost completely occluded shortly after exposure to the gas, and the animal breathed through the mouth with a characteristic gasping reflex.

Immediately upon exposure to chlorine gas the respiration was markedly accelerated, reaching a high maximum within the first hour. From this time until the third hour after gassing the rate was decidedly decreased, but was still far above normal, where it remained, with some fluctuations, for the first fifteen hours. At this time most of the animals had died or were about to die. Chlorine was exceedingly irritant to the respiratory passages; almost immediately upon exposure to the gas, and for many hours thereafter, frothy or stringy saliva dropped from the mouth constantly. The quantity of fluid thus lost to the body might be quite large. Respiration was very difficult and apparently the animal had considerable discomfort but did not appear to be in actual pain.



The immediate effect of phosgene poisoning was a decided lowering in the pulse rate. From an average normal of approximately 90 it was found that the rate would drop to about 75 beats per minute. Many cases were noted in which the rate was less than this. In general the normal rhythm was not reached until the fourth or fifth hour after gassing. However, in some cases, the pulse regained the normal rate during the first hour or so after gassing. The heart might remain at this rate or slightly higher for some hours and then there would be a gradual increase, occasionally a very rapid increase, and in the course of 10 or 12 hours in such cases the heart rate would react close to 150 beats per minute. In the animals which were less seriously affected, the rate in general remained more nearly normal than it did in those which were in a serious condition. The individual observations for the most part showed that the more seriously the animal was affected the higher would be the pulse rate. The pulse rate continued high until the animal died. There was evidence in some cases that the high rate of the heart action resulted in a circulatory failure. In such cases the pulse rate would drop very rapidly until it reached a point which might be considerably below normal and the animal died soon after.

Gassing with chloropicrin caused a sharp break in the pulse rate. It could fall to one-half, or even less, of the normal rate within the first hour. In the hours immediately following, the heart tended to regain its normal beat and in four hours or less the rate in general was normal or above.

In animals that were slightly affected by the gas the pulse rate might remain somewhat above the normal figure. In animals that were seriously affected the pulse rate returned to normal very rapidly and then in a few hours might reach a rate of 180-200 beats per minute, which might continue until death. It appeared, in some cases, that the work put on the heart at this period was too great, and suddenly the rate broke sharply and the animal died within a short time.

Immediately subsequent to exposure of an animal to chlorine gas the pulse rate fell somewhat and then steadily increased until the rate reached 150 or more between the sixth and eighth hour. Thereafter the rate decreased steadily until it usually reached a normal or subnormal value. In animals that were not fatally gassed the pulse might fall sharply within the first four hours and then rapidly rise to a very high figure, which was maintained for many hours (30 or more).


The immediate effect of phosgene poisoning was to cause a break in the temperature. This was apparently a resultant of the lowered pulse rate and the consequent subnormal circulatory efficiency. This break in the temperature averaged about 1°C., so that the normal temperature of approximately 39°C. would fall to 38°C. within one hour after gassing. The average drop was greater in the animals which were not so seriously affected by the gassing. As the pulse rate tended to increase in the hours following gassing, so the temperature of the animal also increased, and it was found that in the great majority of dogs the temperature had reached normal between the fourth and fifth hours after gassing.


In the animals which were not seriously affected by the gas, the temperature hovered around normal, maybe slightly below or slightly above, for the next 7 or 8 hours, and then began a slow drop, so that at the end of the first 24 hours after gassing the average temperature of the dogs was about 38°C. If the animal was withstanding well the effects of the gas the temperature would slowly begin to rise and in the course of the next 21 hours or so reached the normal figure again.

In the animals which were seriously affected by the poisoning the temperature reached normal more quickly after gassing and instead of hovering around normal tended to go above, and cases have been known in which a temperature of above 40oC. was reached within the first 4 or 5 hours after gassing. The average curves showed, however, that a temperature of approximately 39.4o C. was reached in the seriously affected animals during the sixth hour after gassing. Following this the temperature began to break and the more rapidly it fell the more serious was the condition of the animal. A temperature which had fallen to 38°C. or below in from 9 to 10 hours after gassing indicated the death of the animal within less than 24 hours. The seriously affected animals which survived 24 hours after gassing showed on the average at that time a temperature about 3°C. below normal, or 36°C., and the death of the animal could be expected within less than 3 days and generally within 36 hours.

The temperature of animals after gassing with chloropicrin showed a drop which in many cases was very marked and was often as much as 2oC. The average fall was about 1oC within 1 hour after gassing. In the animals that were most seriously affected the temperature continued to fall, and extreme cases were noted in which the temperature during the fourth or fifth hour after gassing was 4oC. below normal. In such cases death usually occurred within 12 hours. In the animals that were less affected the temperature after the initial drop within the first hour after gassing did not vary greatly for the next 3 or 4 hours. Beginning at about that period the temperature began to rise slowly, and in exceptional cases reached normal within 12 hours. However, in most instances the temperature after the fifth hour began to break slowly and normal temperature was not reached for one or two days. The records show that if the temperature during the first 12 hours fell much below 37°C. the animal was in a serious condition.
The temperature changes of animals after exposure to chlorine gas resembled those of the acute deaths induced by chloropicrin. With chlorine, however, the tendency was for the temperature to fall even more profoundly than was the case with chloropicrin just cited. Death usually resulted in such instances in less than 12 hours from the time of gassing.

In animals less seriously affected there was a similar initial fall in temperature which often, however, gradually returned toward the normal, the latter being attained within the first 24 hours.


From a practical viewpoint it is important to know whether an individlual becomes more or less susceptible to the gas by repeated exposure. There seems to be a rather widespread opinion that in man a single exposure to a gas greatly increases susceptibility. This view, however, is founded entirely upon general impressions, and in questions of this kind many psychological factors enter which make a clear-cut definition difficult. On the other hand, it is quite plausible to assume that a mucous membrane once extremely irritated might


be more easily thrown into an abnormal state by a weaker stimulus than would be true for a membrane which had always been normal. To test this point experiments were made with chloropicrin. The animals used had all survived an initial gassing for periods of approximately a month at least and to all appearances were normal.

If the view given above is correct this would indicate that the apparently beneficial effect of previous gassing at relatively high concentrations is due largely to the elimination of the weaker or more susceptible individuals by the first gassing. At any rate, as a result of an investigation with more than 50 animals it was indicated that so far as chlorine was concerned no evidence was obtained of any increased susceptibility. It was shown that susceptible animals were eliminated by the first gassing in proportion to the concentration at which they were gassed, and that the survivors had every chance of recovery from a second gassing at the same concentration. If, however, the second gassing was at a higher concentration, a proportionately increased percentage, could be expected to succumb.

When the problem of regassing with phosgene was investigated one was confronted with an entirely different picture, for phosgene poisoning increased the susceptibility of the animal to this gas.


A problem of fundamental importance in the investigation of the physiological action of inspired gas was whether it, or its decomposition products, actually penetrated the body tissues. Were the poisonous effects of the gas due solely to its action upon the lungs or were they also due in part to absorption into the blood stream and distribution to the body as a whole? A study of the urine would perhaps give an indication in the solution of this problem. This would be true especially with chlorine inasmuch as the changes in chloride excretion might yield a decisive answer to the question.

Selected animals were observed during a six-day normal fasting period, during which time they received water but no food. At the end of each 24-hour period, except the first, the urine was collected and analyzed. The urinary picture thus obtained was taken as an index of the normal metabolism of the subject and afforded a basis for comparison with the data obtained after gassing. After this initial normal period the dogs were fed for a week and then again starved. On the second day of this fasting period they were subjected to the action of the gas, and the urine was collected at the end of each subsequent 24-hour period. This procedure was continued, as a rule, for five days, if the animal survived. The methods of urinary analysis were those commonly employed in scientific investigation.

The general effects of chlorine poisoning on the composition of the urine were as follows: The hydrogen ion concentration was increased, and there was a tendency toward augmented titratable acidity. The "organic acid" figure was markedly increased. The excretion of ammonia, total nitrogen, creatine, uric acid, phosphates, and chlorides was greater than in the normal period. On the fourth or fifth day, the output of creatine, phosphates, and chloride tended to drop below normal. Large urine volumes were frequent and protein was present in the majority of cases. The picture represented was practically the same in all experiments, though the higher concentrations of chlorine yielded more marked effects. Expressed differently, it may be stated that exposure to chlorine gas of varying concentrations caused in the dog a markedly increased protein metabolism. Typical data may be found in Table 20.


TABLE 20.- The influence of chlorine poisoning upon the composition of the urine


In order to determine whether the increased protein metabolism was a secondary manifestation of a profound disturbance in the carbohydrate metabolism, the glycogen content of the liver was estimated by Pfluger's method. Determinations were made eight hours after gassing. The results were compared with control determinations, made on normal dogs that had fasted for corresponding periods of time. The data, as given in Table 21 show that the amount of glycogen in the liver was not materially affected by gassing at high concentrations (2.53 to 2.85 mgm. for half an hour).

TABLE 21.- Chlorine experiments
Since the increased elimination of nitrogeneous substances was not accom- panied by a complete removal of the glycogen in the liver and therefore could not be attributed to utilization of protein in place of carbohydrates, it was indicative of destructive processes within the tissues. Possibly it was to be associated with autolytic decomposition in the lungs. In harmony with this idea was the fact that the maximum output of nitrogen fell on the second day after gassing, which was synchronous with a crisis in the lung condition.

The increased acidity of the urine and the augmented excretion of ammonia, acid phosphates, and "organic acid" all indicated acidosis, a condition which was intensified by the disturbance in protein metabolism. That the acidosis was not primarily dependent upon the augmented protein metabolism was indicated by the conditions that obtained in dogs gassed at low concentrations.

The chloride picture characteristic of severe gassing showed chloride retention during the first 24 hours, followed by increased elimination, the maximum output usually falling on the second day. This condition was intimately associated with the concentration of the blood, the period of retention being synchronous with the period of blood concentration and the subsequent increased excretion occurring when the blood returned to a more dilute condition. The increased chloride output in the urine probably can not be accepted as evidence of chlorine absorption during gassing.c This was corroborated by the data from the chloride content of the blood and tissues. Moreover, the urine analyses failed to show any proportionality between the concentration of the gas administered and the chloride increment eliminated in the tissues.

c In this connection search was made for free chlorine in the blood and plasma of dogs before and after gassing. The inorganic chlorides were first determined by the method of McLean and Van Slyke. Samples of blood were also ashed with CaO and Na2CO3 and the chlorides determined in the fusion mixture. This procedure obviously would convert any free or organically combined chlorine to chloride and permit its determination as such. No significant difference was found between the two sets of analytical results. Also quantitative examinations of the protein-free filtrate from the plasma or blood for free chlorine, by means of the reaction with KI and starch paste, were equally negative. If free chlorine occurs in the blood of gassed animals it is in traces too small to admit of detection by the methods employed. It is obvious, therefore, that chlorine in appreciable amounts does not exist in the blood other than as ionized chloride, and accordingly that direct chlorine absorption by the blood is not a significant factor in gas poisoning.


Dogs gassed at a moderate concentration of phosgene, 60 to 70 parts per million of air (0.26 to 0.31 mgm. per liter), showed the following metabolic changes: The nitrogen metabolism was increased during the second 24 hours after gassing, paralleled by a very high excretion of creatine during that period. The chloride output was very high in the first 24 hours, low in the second, and high in the third, gradually returning to normal on the fifth day. The phosphate output was very high during the first 24 hours, then gradually decreased until it reached the normal value during the latter part of the experimental period. The sum of the daily phosphate excretion during the period was about the same as that of the normal period. Throughout the interval there was no evidence of diuresis, and the hydrogen-ion concentration was not affected. The "organic acids" were unchanged and remained practically constant for both the normal and the experimental periods.

Dogs gassed at a higher concentration of phosgene, 90 to 100 parts per million of air (0.40 to 0.45 mgm. per liter), presented a picture similar to those gassed at a moderate concentration, except that the chloride elimination was not as high during the first day. With lower concentrations of phosgene, 40 to 50 parts per million of air (0.17 to 0.21 mgm. per liter), the typical changes observed at moderate concentrations were present, although the extent of alterations was less marked. Tests for various abnormal constituents of the urine gave no evidence of a pathological condition.

With chloropicrin the nitrogen metabolism was increased on the second day, although in some cases it began to increase on the first day, resulting in an augmented output of total nitrogen, ammonia nitrogen, uric acid nitrogen, and creatine nitrogen. Chloride output was seldom above the normal figure. This was especially true in dogs moderately gassed. (Table 22.)

TABLE 22.- The influence of chloropicrin poisoning upon the composition of the urine


TABLE 22.- The influence of chloropicrin poisoning upon the composition of the urine - Continued.


The phosphate elimination was increased greatly during the first 24 hours, often being double the normal figure. In most dogs that died within a day or two after gassing the phosphate output during the first 24 hours was only slightly above normal, and when this was found to be the case it was quite certain the animal would not survive. The titratable acidity between PH 4.9 and PH 7.4 ran parallel with the phosphate output, showing that this was a simple titration of the "buffer" reaction of the phosphates. The volume of the urine, hydrogen ion concentration, and "organic acids" showed little or no change as a result of gassing.

Gassings at a low concentration showed in general, the same picture as those of a moderate concentration, except that the changes were less marked. Generally the chloride output was a little higher. Gassings at a high concentration showed a picture similar to that at a moderate concentration, but the effects were somewhat prolonged.

Kidney efficiency tests were run on some dogs inasmuch as it had been asserted that chloropicrin might injure the kidneys. In the cases tested no decreased renal function could be detected.

A comparison of the excretion under the influence of the three gases showed that in all three instances nitrogenous metabolism was definitely increased, the various partitions of nitrogen running more or less parallel curves. In certain instances the increased output was most evident on the first day subsequent to gassing, in other cases the second day showed the greatest excretion. Chloride elimination was very markedly increased in some instances, as in the case of chlorine gassing, or only of slight significance as with chloropicrin, or again it might assume a widely divergent curve as with phosgene. Chloride excretion was undoubtedly linked with the changes in the development of edema; the chloride output, therefore, will be discussed more fully in connection with the development of pulmonary edema (pp. 333-342).

Acidosis might or might not have been present as indicated by changes in the hydrogen ion concentration, titratable acidity, and "organic acid" figures. These results make it evident that there was, therefore, no essential relationship between the increased nitrogen output and acidosis.

Creatine excretion was quite prominent but seemed to follow no definite course. Apparently it was not associated with lack of carbohydrates,1 as for example with chlorine poisoning, nor could its appearance in the urine be ascribed to a condition of acidosis 2 since in neither phosgene nor chloropicrin poisoning was there any indication of such a state. It was possible, however, that it might have been due to tissue changes induced in the lungs whereby disintegration occurred with the formation and subsequent elimination of creatine. By such a process both creatine excretion and increased nitrogenous metabolism would be explained, although it must be admitted that the reactions involved are by no means clear. It is quite significant that a second exposure to a gas rendered an animal neither more nor less susceptible as judged by the influence upon nitrogenous metabolism. So far as investigated there was little evidence that the lethal gases, employing chlorine as an example, were absorbed by the blood stream. At most only the merest traces were absorbed. The damage to the organism was therefore localized upon the respiratory tract.



Even a slight experience with gas poisoning led to the recognition that changes in blood concentration must occur. The evidences for such an impression were not prominent in the early stages, but as time passed it became quite apparent that the blood assumed a sticky, concentrated consistency, attempts to draw blood from a vein, for example, being attended with great difficulty. The character of the blood at this period exerted a definite detrimental influence upon the rate of circulation, reacting in time to impede the heart action, and later on even to interfere with the proper blood supply to the tissues. Under these conditions the tissues consequently suffered, and normal metabolism, therefore, was undoubtedly distinctly altered.

The changes in blood concentration have been studied extensively in this investigation, since it has been assumed that such alterations were quite suffi-

CHART VI.- Changes in total solids of blood after phosgene gassing, showing characteristic differences in the three gas concentrations selected

cient to explain many of the phenomena associated with gas poisoning. From the fact that observations have been made most intensively with phosgene, the results obtained with this gas will be presented first.

An inspection of Chart VI shows that at first the blood usually contained less solid matter than normally. This condition was maintained for several hours (stage 1). Later concentration began and rapidly assumed a maximum (stage 2), after which there was a gradual return (stage 3) to the normal level. The results showed very characteristic differences between the three gas concentrations selected. At 80 parts per million and above there was a very rapid recovery from the preliminary dilution, the normal being regained within 4 to 5 hours. The succeeding concentration was correspondingly rapid. reaching a maximum at 12 to 14 hours and was at a level 18 to 20 percent


above the normal. Practically all of these dogs died during the second stage so that the third stage was not represented at this concentration.

With 60 to 80 parts of phosgene per million of air, stage 1 was prolonged beyond that at the higher concentration and the dilution persisted for 4 to 6 hours. The succeeding increase in the total solids also developed more slowly and was not so great, being only 16 to 18 percent at 18 hours. More of these dogs survived the acute period and stage 3 appeared in the curve. The return to normal was practically complete by the forty-fifth hour. At the lowest concentration studies, 40 to 60 parts per million, the changes were still less in degree. The dilution period extended over 6 to 8 hours and the maximum concentration of 12 to 14 percent was not gained until 21 to 22

CHART VII.- Changes in total solids of bloo after phosgene gassing, showing characteristic differences in the dogs that died acutely and those that survived the acute period.

hours after exposure. Stage 3 was very similar to that with the intermediate concentration. Compared to the other two series, but a very few dogs in this group died.

A study of the average results outlined above would indicate a very distinct relationship between the phosgene concentration and both the rate and degree of change in the blood. As has been mentioned before, practically all the dogs in the first group died acutely, while but very few died in the last group. The dogs in the second series were about equally divided between those that died acutely and those that survived the acute period.

Careful analysis of these results indicates (Chart VII) that within each group the changes in the blood concentration showed characteristic differences, depending upon the ultimate fate of the animal. In dogs that died acutely the period of dilution was short and the normal was regained within


three to five hours. This was followed by an extremely rapid concentration that reached a maximum of 22 to 23 percent over the normal 15 to 16 hours after gassing. All these dogs died during this period of greatest blood concentration.

The survivals show a different picture (Chart VII). The initial period of the reaction was slower, up to the sixth to tenth hour. Owing to the marked variation among individuals no distinction could surely be made between the acute deaths and survivals on the basis of the degree of dilution during this primary stage. The difference in time was marked. The great difference, however, came in the second stage, for the dog that ultimately recovered showed a very much slower development of the blood change and, in addi- tion, the concentration was not nearly so marked, 12 to 15 percent 22 to 24 hours after gassing. Following this came the stage of recovery, which has alreadv been discussed.

These two types of classification with the reaction characteristic of each hold throughout the range of gas concentration studied. On the basis of this generalization it may be pointed out that the apparent correlation between response and phosgene concentration in the previous experimental series was due to the relative predominance of two separate types of reaction and not to the gradual change in the type of reaction by individual animals.

From these graphs it is quite apparent that three distinct stages may be recognized as occurring in the blood solids subsequent to phosgene poisoning.

The first stage is apparently one of dilution of the blood as evidenced by a decrease in the blood solids. This dilution is greatest one to three hours after gassing and the total solids have returned to normal by the fifth to eighth hour. The cause of this sudden decrease in blood solids is not entirely clear.

The second stage is one of blood concentration. The total solids of the blood increase rapidly to a value far above the normal and remain stationary at this level for several hours. In the dogs gassed at 90 parts per million the average value for the total solids increased up to a maximum of 25 percent 10 hours after gassing and remained at approximately that level until the death of the dog. In those animals gassed around 70 parts per million the average value does not reach a maximum until some 171 hours after gassing, and even at this time the value is lower (23 percent) than in the case of the higher gas concentration. The speed of blood concentration and the degree are both greater with the higher gas concentration.

The third stage
marks the gradual return of the blood solids to the normal level.

Owing to the greater density of the red corpuscles of the blood as compared with the plasma any change in the relative amounts of corpuscles and plasma will cause a corresponding change in the total solids. The observed changes in the blood solids, therefore, might have been due either to an increase in the plasma volume during the first stage, followed in the second stage by a decrease and with no change in the erythrocytes; or else to the withdrawal of erythro- cytes from the circulation in the first stage, and a later reintroduction. According to Lamson 3 such changes are possible through the mediation of the blood sinuses of the liver. If the first of these possibilities is correct then the blood volume should be increased during the first phase and decreased later when concentration occurs. The reverse would be true in the second case.


Eyster 4 reports that radiographs taken during the early stages of phosgene poisoning show a dilated heart, but without an increased plasma volume. On the other hand, in the second stage the heart is markedly decreased in size.

From the foregoing it is evident that the changes in the concentration of the blood as determined by total solids gives one a method of following the condition of the animal in this respect.

The changes in concentration might equally well be followed by the determination of the hemoglobin. Hemoglobin determination is much more rapid, less blood is necessary, and the method is even more accurate than the more cumbersome total solid determination. Accordingly, a comparison has been made between the hemoglobin content and the total solids of the blood. From this it may be concluded that the two curves are similar but not parallel, and that the degree of change in the hemoglobin at all periods after phosgene poisoning is much greater than is true for the total solids. The hemoglobin, therefore, is a much more delicate indicator of the animal's condition than is the total solids. It has been employed to estimate the changes in blood concentration.

A series of hemoglobin determinations was made upon a comparatively large number of dogs gassed with phosgene at concentrations varying from 41 to 80 parts per million of air (0.17 to 0.35 mgm. per liter). On the basis of these studies the following different types of individual reactions, together with the probable fate of the animal, may be outlined:

Type 1.-The reaction of the animal was slight, there being a variable degree of dilution followed by a return to the normal hemoglobin value. Stage 1 alone was apparent in this case, the concentration being absent. The recovery of the animal was uninterrupted after the passage of the dilution.
Type I.-This type followed the usual stages of blood change, all three stages being present. The concentration in the second stage was relatively moderate, varying up to 140 percent of the normal. Under these conditions the animals recovered.

Type III.-This type was differentiated from the second type by the degree of concentration of the blood. Concentration of over 140 percent was usually fatal. Death usually occurred in the second stage.

The fate of the animal was dependent on two factors: (1) The degree of the concentration of the blood, and accordingly the extent of the edema, and (2) the individual resistance. In general, however, it was found that a concentration of 140 per cent marked the mean between the two conditions. Hemoglobin readings above this indicated the probable death of the animal; below this, the recovery.

Type IV- This type was characterized by the absence of any dilution following gassing. The concentration appeared immediately, or within the first two or three hours. The blood changes in this case were rapid and extreme, and were usually followed by early death.

Earlier in this chapter it was stated that on the basis of the changes in the total solids the animals could be separated into two groups, one of acute deaths, and a second of survivals. These two classes are practically synonymous with the reaction Types II and III, as outlined above. On the basis of the greater number of animals studied it has been possible, in addition, to characterize Types I and IV.


These are the changes in the blood as indicated by the hemoglobin readings. The individual type of reaction seems to be characteristic of no particular gas concentration, as examples of each type were found at all concentrations studied. With the increasing toxicity of the higher gas concentration, greater proportions of the animals showed reaction in order of Types III and IV, while at the lower concentration Types I and II predominated. The number of animals studied at each gas concentration was not great enough to permit of a percentage analysis. In the discussion of the result at each gas concentration, however, this factor of the different types of reaction must be kept in mind. In Table 23 are given data showing the average figures obtained when the results are divided into (1) recoveries, (2) acute deaths, and (3) delayed deaths.
TABLE 23.- Hemoglobin changes in the blood in phosgene poisoning

Dogs gassed at 41 to 50 and 51 to 60 parts phosgene per million of air were affected about equally, i. e., 72 percent survivals and 28 percent deaths, and 69 per cent survivals and 31 percent deaths, respectively. The average hemoglobin picture shown by these two concentrations was almost identical. The first stage showed a dilution of 90 percent of the normal hemoglobin in the second hour, this being followed by a gradual increase until the normal concen- tration was reached between the seventh and eighth hours. There was then a gradual concentration of the blood until the maximum of 125 per cent hemoglobin was reached in the nineteenth hour. The blood began then to dilute and reached its normal concentration about the thirty-sixth hour.

As the concentration was increased (61 to 70) the toxicitv was markedly changed: Recoveries 45 percent, deaths 55 percent. The time factor in the average hemoglobin curve was slightly changed, but the curve as a whole was not altered. The blood diluted to 87 percent of the normal concentration one hour after gassing and remained constant until the fifth hour. The blood reached its normal concentration in the ninth hour and continued to concentrate until a maximum hemoglobin concentration of 120 percent was reached in the seventeenth hour. The return to normal then followed.


At the concentration of 71 to 80 parts per million phosgene the recoveries totaled 37 percent and deaths 63 per cent. Following gassing there was a minimum dilution of 81 percent after two hours, this being followed by a gradual increase until the tenth hour. Between the ninth and eleventh hours there was a marked increase of 20 percent - a jump in two hours from below normal to its maximum concentration (116 percent). The hemoglobin curve remained constant until about the twenty-fifth hour and then began its return to normal.

From the above discussion, the following facts are evident:
1. The blood was most dilute between the first and third hours after gassing.
2. The return to normal after the dilution took place sooner at low than at high concentrations.
3. The time of maximum concentration of the blood during stage 2 depended on the phosgene concentration; the higher the gassing, the sooner is this point reached.
4. The average maximum concentration for recovered animals was about 120 percent of normal hemoglobin.


The acute deaths have been classified as those dogs dying within 72 hours after gassing. The number of dogs dying within a few hours was relatively small, so that an average curve was not indicative of great accuracy. The individual type of hemoglobin curve may be discussed to better advantage.

The most common type of curve was one in which there was a significant dilution followed by a very marked rise in hemoglobin. The time when the concentration began varied from 3 to 10 hours after gassing and was not dependent on the concentration of the gas or the dilution of the blood during stage 1. Another type which invariably proved fatal in phosgene poisoning was an immediate concentration of the blood. The faster the blood concen- trated, so much the sooner did the animal die. In certain instances, a dilated heart caused an acute or delayed death without any sign of significant blood concentration. This, however, was seldom seen.

With chlorine too few experiments were made to warrant more than the most general statements relative to blood concentration changes. The data, however, allow one to be certain that the course of blood concentration alterations, as determined by estimation of the total solids, paralleled the curve obtained by determination of the hemoglobin values. These data also indicate quite clearly that the period of blood dilution, that is,--stage 1 in phosgene poisoning-was either very slight and short lived or lacking altogether, and that the significant feature relative to the blood changes under discussion in chlorine gassing was the almost immediate tendency for concentration and the rapid development of this to a high maximum. (Chart VIII.)

The striking feature in blood changes as indicated by total solid determination after chloropicrin gassing was the usual absence of the first or dilution stage which characterized phosgene. In surviving dogs the blood steadily concentrated and attained its maximum about 10 hours after gassing. Very gradually the blood then became less concentrated until about the fortieth hour it had usually reached its normal value, and became then, for a time, less concentrated than normal.


On the other hand, 20 of the 58 chloropicrin dogs studied showed a dilution period corresponding to the first stage in phosgene. This lasted for a variable interval, but was usually less than two hours, often less than one-half hour in length. In all but three cases the maximum dlution value was found in blood collected within the first hour after gassing. No animal showing this dilution died. The only two animals able to survive gassing above 1.09 mgm. per liter exhibited this dilution stage.

The maximum concentration was about the same as in phosgene, averaging 113 percent of normal. The extent of concentration was to a certain degree dependent on the gassing strength, as it will be seen that the blood of animals gassed below 0.80 mgm. per liter failed to become as concentrated as in animals gassed above this value.

CHART VIII .- Comparison of the changes in total blood solids of dogs gassed with phosgene and those gassed with chlorine

Since, from the standpoint of treatment, the time relations of blood concentration are important, a comparison is made in the table between the effects of phosgene and chloropicrin:


The concentration was attained earlier in chloropicrin, but the variation in time was wide. In dogs which suffered early death from chloropicrin gassing, the rise to a maximum concentration value was swift. In several cases a concentration above 130 percent normal was attained within five hours. After reaching the maximum concentration the animal usually, though not always, died within a short time, (five hours). A comparison of the relationship between the total solid curve and that of the hemoglobin may be seen by inspection of Chart IX.

If one draws characteristic curves of blood changes induced by the three gases, the diagram in Chart X would be the result. The most striking feature of the blood in relation to exposure to the lethal war gases is the marked change

CHART IX.- Comparison of total solids and hemoglobin after chloropicrin gassing

of concentration, which varies characteristically both in degree and time with the different gases. The significance which it is believed attaches to this phenomenon will be discussed in succeeding pages.


A brief study was carried through of the influence of the lethal gases upon the red and white cells of the blood and in certain instances comparison was made of the changes in the red cells and the hemoglobin.

At first observations were made with chlorine, relative to the influence of gassing upon the red cells, over an extended period; that is, cells were counted on successive days. In later work the red cell estimations were made more frequently, at intervals of hours instead of days (Chart XI). From the


data at hand it is indicated that almost immediately after chlorine gassing there occurred a characteristic rise to a high maximum of both red cells and hemoglobin. The graphs, which may be plotted from the data, more or less parallel each other. The increase in the red cell count, however, was usually somewhat greater than that of the hemoglobin value. The approximate parallelism for these two elements of the blood leads to the conclusion that the increase was apparent rather than actual. Stated differently, the apparent changes in the red cell count and the hemoglobin figures are to be referred to changes in the concentration of the blood and can not be accepted as evidence for the intrusion of new cells into the blood stream.

CHART X.- Comparison of the characteristic changes of blood solids induced by chlorine, phosgene and chloropicrin gases

  With phosgene a similar conclusion must be drawn withl respect to alterations in the number of red cells. Shortly after gassing the cells diminished in number and later rose far above the normal value, in harmony with the observed changes in blood concentration, namely, a period of dilution shortly subsequent to exposure to the gas followed by an interval when the blood becomes highly concentrated.

Chloropicrin gassing produced changes in the red cell content and hemoglobin values of the blood somewhat analogous with those induced by chlorine and in entire accord with what might be anticipated by one with a knowledge of the alterations in blood concentration induced bychioropicrin. (Chart XII.)

The data for white cell counts of the blood are incomplete in that no determinations were made with either phosgene or chloropicrin. However, a fairly extensive study was mtade with chlorine, the results of which follow.


Leucocyte counts were made on animals gassed with chlorine in concentrations from 0.33 mgm. to 6.32 mgm. per liter, with variable intermediate concentrations.


Animals gassed at extremely low concentrations (i. e., 0.18 mgm. per liter) exhibited a slight leucocytosis within three to five hours after gassing, which was followed by a return to normal almost immediately, except in cases where the animal developed a slight bronchitis. The counts were not followed beyond the third day subsequent to exposure to the gas. (Table 24.)

CHART XI.- Comparison of erythrocytes and hemoglobin content of blood after chlorine gassing

TABLE 24. – The influence of chlorine gassing upon the leucocyte count


TABLE 24.- The influence of chlorine gassing upon the leucocyte count-Continued

CHART XII.- Comparison of erythrocytes and hemoglobin content of blood after chloropicrin gassing


A detailed study of the changes in leueocytes was made in a series of six dogs gassed at concentrations varying from 1.23 to 2.21 mgm. per liter. Three of the dogs died, showing upon autopsy different stages of pneumonia, while two survived, passing through a stage indicating bronchopneumonia, considerable time elapsing before their full recovery.

The extent of the leucocytosis varied greatly, but showed a typical picture in any case. One dog showed very few symptoms and gave the same picture as; a dog, gassed at very low concentration, i. e., a slight leucocytosis and a return to normal the next day. The remaining dogs that recovered with pulmonary complications developed a moderate leucocytosis, which continued for several weeks, followed by a return to normal on complete recovery.

The fatal cases in this series showed typical curves which corresponded to the condition of animals gassed at very high concentrations. There was one type which developed an extreme leucocytosis, followed by a fall in count before death, and another in which not even a moderate leucocytosis appeared. In the latter, autopsies revealed severe cases of purulent bronchopneumonia. It may be assumed, therefore, that the development of continued moderate leucocytosis (about 200 percent) was essential for the protection of the organism in cases where animals had been gassed at a moderately high concentration.


Animals of the last series were gassed at very high concentrations, far above the lethal dose. This procedure was followed in order to see if it were possible to diagnose a fatal case of chlorine poisoning from the leucocyte count. Four distinct types of curves were observed: (a) In one case an extremely high leucocytosis was followed by death a few hours after gassing, and in another a gradual fall in count for several days, after which death resulted. (b) A moderate leucocytosis followed by a sudden drop in count on the day after gassing. (c) A failure of the organism to develop leucocytosis, in which case death resulted in about three hours after gassing. (d) A slow development of a leucocytosis followed by marked fluctuations in count, death occurring within 10 hours.


A partial study of the differential picture showed that leucocytosis was caused solely by an increase in polymorphonuclear cells, the lymphocytes and mononuclear cells playing no part. The eosinophiles disappeared from the circulation for a short time several hours after gassing. The data presented above may be restated from the standpoint of whether or not the dog recovered from the chlorine poisoning. Dogs that recovered showed two types of curves: (a) After mild gassing a slight leucocytosis (100 percent) followed by quick return to normal; (b) after lethal concentrations of chlorine, a moderate leucocytosis (300 per cent) persisting for several weeks.

Dogs which exhibited the following leucocytotic condition invariably died: (a) A leucocytosis; (b) an unstable fluctuation in the leucocyte count during the first few hours after gassing; (c) a moderate leucocytosis followed by a marked drop (acute and chronic cases); (d) an extreme leucocytosis during the first few hours after gassing.


Pulmonary edema is a very prominent feature of the effects of the lethal war gases on the animal organism. To its development attaches great significance in any explanation of the detrimental influence of a gas. Equally important is a consideration of the subject when attempts are made to define the cause of death in the circumstances under discussion.
The lethal war gases are all substances eminently irritant to living tissues, and it must be accepted that the irritation produced by a gas is the initial step in the development of edema. In response to the first irritative stimulus, tissue fluid finds its way to the injured area in an apparent attempt toward


repair or alleviation of the injury. It is conceivable that if damage to the tissue is only slight such a procedure would result in the passage to the damaged area of only a small quantity of tissue fluid. According to this view the degree of response with respect to the local deposition of tissue fluid would be in direct ratio to the extent of injury. On the other hand, it is equally plausible to assume that this reaction may reach a breaking point at a certain degree of stimulation whereby the whole mechanism governing the exudation of tissue fluid is thrown out of control so that the response to the stimulation becomes overwhelming. Under these conditions a reaction which in its initial function may be regarded as beneficent eventually becomes a direct menace to con- tinued existence on the part of the mechanism as a whole merely by interposing mechanical difficulties in the way of respiration and circulation.

It is not proposed in this place to discuss in detail the underlying principles of edema production. Rather an endeavor will be made to correlate so far as possible various observations that have been carried through in this investigation with the development of pulmonary edema. At this time, therefore, attention is called to the development of edema of the lungs in its time relations; the correlation of pulmonary edema with changes in blood concentration; the association of edema with chloride and fluid exchange in tissues and the blood, and the relation of edema to vascular permeability.


In connection with the determination of the toxicity of phosgene for dogs a relationship was observed between the time of death and the concentration of the gas to which the animal had been exposed. In general the greater the concentration the sooner the occurrence of death. Accordingly, in this series, attention was confined to a concentration somewhat below the lethal (70 parts per million; 0.31 mgm. per liter), and to one somewhat above lethal concentration (90 parts per million; 0.40 mgm. per liter). In all cases a standard time of exposure for 30 minutes was used.

The dogs were killed by strychnine injection at intervals after exposure and samples of tissue were taken from different parts of the lung. An effort was made to secure as composite a sample as possible and to reduce to a minimum the loss of exuding edema fluid during sampling. There was always slight loss, particularly with very wet lungs, so that the results mav not have been quite as accurate as when the entire lung was dried. The error, however. was negative rather than positive and the degree not sufficient to compensate for the difficulties of analysis by the latter method. The samples for the determination of the total solids in the blood were drawn at regular intervals by needle from the jugular vein. All samples were carefully dried to constant weight at 105o

The total solids of the lung (Charts XIII and XIV) showed a rapid and extreme decrease indicating the production of an intense edema of the most marked type. This was most noticeable in dogs Nos. 481 and 479, killed at eight and nine hours after gassing. In these two animals the total solids of the lung fell to 8.6 and 7.1 percent from a normal value of about 21.4 percent. Assuming that the decrease in total solids was due to the influx of water alone into the lung these values would represent the influx of 150 and 200 cubic centimeters, respectively, of water per 100 grams of original tissue, a truly


enormous increase. In view of the fact that the edema fluid contained most of the constituents of plasma and that no allowance was made for these, the actual inflow wvould have been, therefore, greater than that here calculated.

The experimental series was small in view of the variation among indi- vidual animals, but, a careful examination of the curves (for example, Chart XIV) shows that in general the animals gassed at a concentration of 90 parts per million (0.40 mgm. per liter) showed a given total solid content 4 to 6 hours before those gassed at a concentration of 70 parts per million (0.31 mgm. per liter). At a given time the total solid content was 1½ to 2½ percent lower in the dog gassed at the higher concentration.
CHART XIII.- Changes in the chlorides and total solids of the lungs after phosgene gassing

A maximum degree of change was present in both sets of analytical results from 10 to 25 hours after gassing. After this time there was a more gradual return to the normal lung condition.   

Analysis of lungs of dogs subjected to chlorine gas showed an immediate influx of water to a marked degree. The water of the lungs gradually subsided if the animal survived a sufficiently long period.

With chloropicrin (Chart XV) the water content changes of the lungs were very significant.

The changes in the lung indicated the rapid production of a marked pulmonary edema and its gradual subsidence with all three gases. The rapidity of the production of this edema depended in part, at least, upon the concentration of the gas employed and there were indications that there was a direct relationship between gas concentration and the degree of edema.



In Chart XIV a comparison is made between the rate and extent of edema production and the changes in the blood solids in phosgene poisoning. The courses of the two processes are fairly synchronous, the development of edema corresponding with a fair degree of accuracy to the concentration of the blood. The only explanation for the initial dilution of the blood is that at first fluid passed more rapidly into the blood than it could pass from the blood into the lungs. In general the blood returned more rapidly to the normal level than was true for the lungs. In the latter instance, however, there was a possibility of interfering secondary factors, such as pneumonia, which might complicate the matter.

CHART XIV.- Changes in the total solids of the blood and lungs after phosgene gassing

From the data of fluid changes in the lungs, and blood concentration alterations respectively, what has been concluded relative to phosgene applies equally well to chlorine. To even a greater extent is this true, for chloropicrin (Chart XV.)

From the data represented it may be concluded that for the production of edena of the lungs induced by all three gases, fluid is drawn from the blood. Moreover, since it has been shown that the hemoglobin and total solids have similar types of curves the estimation of hemoglobin may be employed to follow the course of blood concentration and hence in general to act as a criterion of the derelopinent, or staqe, of pulmonary edema.



In any study of the production of edema the question of possible disturbances in the salt relationship is at once raised. In an effort to secure a partial elucidation of this problem a study was made of the chloride content of the blood and lungs of dogs gassed with phosgene. No study was made in this connection with chloropicrin.

In this investigation attention was confined at first to concentrations of 90 and 70 parts of phosgene per million of air, i. e., slightly above and slightly below the lethal concentration. Blood chlorides were estimated by the method of McLean and Van Slyke. The chloride content of the lungs was determined in the dried tissue used for the study of the total solids. The method of analysis was a modification of the McLean and Van Slyke procedure.

CHART XV.- Water content of lung tissue after chloropicrin and phosgene gassing
In normal starving dogs the chloride content of the blood plasma was approximately constant from day to day, but there was a wide variation among individuals. The results have been expressed in terms of the percentage of the normal value. Since it was found that the plasma and whole blood chlorides underwent parallel changes, only the plasma chlorides were determined. More uniform alterations were obtained when the animals were starved for 48 hours before gassing than when inanition was for a period of 24 hours only. Owing to the small experimental series no distinction was made between animals starved for different periods or gassed at the different concentrations, the discussion being confined to the average of the entire experimental series. The results of these observations are expressed graphically in Chart XVI.


There was practically no change in the plasma chlorides during the first four hours after gassing. Between the fifth and sixth hours, however, there was a marked and rapid drop in the blood chlorides. This drop reached a minimum at about the twentieth hour, and from then on the blood chlorides showed a gradual increase during the first two days after gassing until about the fiftieth hour when they gradually returned to normal. During the period of blood dilution there was practically no change in the blood chlorides, indicating that the diluting fluid must have been isotonic with blood plasma. The blood chlorides, however, dropped sharply at about the time the concentration of the blood first became marked.

It has already been shown that blood concentration was due to the passage of fluid from the blood into the lungs. Examination of the lungs, furthermore,

CHART XVI.- Relation between the changes of plasma chlorides and lung solids after phosgene gassing

showed a rapid increase in the chloride content following gassing (Chart XVII). There was a latent period in this inflow of chlorides extending over the first three or four hours, followed by a very rapid increase. Maximum values were obtained by the tenth hour and were maintained during the rest of the first day following gassing. After this period the chlorides left the lung and the chloride content gradually returned to normal. Complete data are not available, but the process of chloride disappearance from the lung was well advanced by the fiftieth hour and the normal was regained soon after.

From these results it becomes evident that the entrance of chlorides into the lungs determines the blood values. The amount of blood chlorides (Chart XVI) was unaltered until retention developed in the lungs. Following the


development of this lung condition the blood chlorides dropped to a minimal value and were so maintained until the fiftieth hour when a return to normal commenced. This corresponded to the passage of the acute pulmonary condition.

In animals dying from acute edema and autopsied immediately, samples of the pulmonary exudate were obtained by removal of the entire lungs and collection of uncontaminated fluid as it ran from the trachea. The fluid collected in this way was clear, straw colored, and occasionally contained a few erythrocytes. It clotted on standing. Determination of the chloride content of this fluid showed essentially the same value obtained from a simultaneous sample of blood plasma. (Table 25.) This indicated that there was

CHART XVII.- Changes in the chlorides of lung tissues and blood plasma after phosgene gassing

complete permeability of the lung capillaries for salts and that the pulmonary exudate and blood plasma were in complete equilibrium with reference to their salt content even if not exactly identical in composition.

TABLE 25.- Chloride content of blood plasma and pulmonary exudate after phosgene gassing


The relationship between the blood and urine chlorides was not entirely clear. The high urinary excretion of chloride during the fourth to seventh hour after gassing showed no direct connection with either the blood or pulmonary changes. During the period of chloride retention in the lung and of low blood content the urine chlorides were well below normal and only in- creased on the third day after gassing, when the chlorides in the lung were liberated and the blood content again rose above the threshold of kidney excretion.


With phosgene and chloropicrin the data available were too incomplete to warrant a decisive inference relative to the part played by tissue chlorides in the production of pulmonary edema. Hence these data are not included in the present discussion. With chlorine gassing, however, the liver, and to a smaller extent the muscles, showed a distinct tendency toward a decrease in the chloride content from 4 to 10 hours after gassing followed by a return to normal between 30 and 40 hours later. The change in the H2O content of the liver was less marked than the chloride content, but the two changes tended to be parallel. It is realized that the evidence presented on this point is not entirely conclusive and it is included merely for the purpose of indi- cating the probable transport of chloride and fluid from the tissues to the blood.


A satisfactory explanation for the concentration of the blood during the production of edema is difficult. At least two hypotheses may be formulated: (a) Fluid is taken from the tissues to a maximum degree by the blood, whence it localizes in the lungs, the final blood concentration being caused by the inability of the tissues to supply further fluid demands made upon them by the blood. It is conceivable that extraction of fluid from a tissue, like muscle, can proceed to a limited extent only if normal processes are to obtain. (b) The hypothesis is plausible that blood concentration is caused by mere ex- traction of fluid from the blood, the exit of fluid from the blood producing edema not only of the lungs but of other tissues as well. If such a view is pertinent the degree of edema in tissues other than the lungs must be slight, inasmuch as there is no visible evidence of such a condition.

Analyses were made of the total solid content of tissues of dogs gassed with chlorine, phosgene, and chloropicrin. It was obvious that in order to obtain conclusive evidence in support of either of the above hypotheses a large number of determinations had to be made, owing to the individual variation of dogs with respect to the total solid content of the tissues. The data are insufficient to draw dogmatic conclusions, but are ample to indicate that the second hypothesis, namely, that there was a general edema of the tissues, does not hold. In other words, edema was not general. On the other hand, there is evidence for the view that water was drawn from the muscles.

As a tentative hypothesis it may be accepted, therefore, that during the development of pulmonary edema fluid was drawn from the other tissues--perhaps specitically from muscle tissue.



A discussion of edema would be incomplete without reference to a possible change in the permeability of the blood vessels, either general or local, since in a theoretical consideration of the production of edema it is generally assumed that alterations in vascular permeability may be a significant factor.

In the experiments to be reported, dogs were infused with sodium-chloride solutions before and after gassing, and changes in blood volume were measured at short intervals after infusion. This method gave the rate of return to the normal of the augmented blood volume after infusion; in other words, the rate of disappearance of the infused fluid was determined, and an indication was obtained of the comparative permeability of the blood vessels.5

Considerable careful work was first done to determine the applicability of the determination of hemoglobin to the measurement of blood volume. Samples taken from ear vessels were absolutely worthless unless a clean cut was made with the lancet, resulting in a copious flow of blood which appeared without manipulation or rubbing. Into the jugular vein of normal dogs was infused physiological saline solution, at 38o so that a volume equal to 1 percent of the body weight was injected in approximately one minute. Samples were taken from the ear before infusion and at one minute intervals after infusion until the blood volume returned to that indicated by the hemoglobin before infusion. Two-hundredths of a cubic centimeter of blood was measured by a Sahli pipette and delivered into 6 cubic centimeters of weak ammonia water. Coal gas was then passed through until the hemoglobin was changed to CO-hemoglobin, when the color was compared to a standard 1 percent solution of CO-hemoglobin in an Autenrieth calorimeter.

With gassed dogs the procedure was the same with the exception that blood to the extent of 1 percent of the body weight was withdrawn one hour after gassing, as in the standard treatment.d The infusion was made five hours after gassing, when the blood usually had concentrated above normal. A comparison of the time for the blood volume to return to normal after infusion in the normal dog with the corresponding time in the gassed dog, gave an indication of comparative permeability.

The results obtained were definite. The time for the infusion fluid to disappear varied from 0 to 21 minutes in normal dogs and from 8 to 33 minutes in gassed dogs. In all dogs save one, the time for the infusion fluid to disappear was longer after gassing than in the normal dog. The time for infusion fluid to disappear in normal dogs varied; i. e., it was an individual characteristic and the decrease in permeability after gassing likewise varied with different dogs. It should be emphasized that there was no evidence of increased permeability, with a single exception, and a very definite indication that the permeability of the vessels was somewhat decreased during the stage of phosgene poisoning studied.

From the foregoing considerations it may be accepted that the development of edema as a result of the action of the lethal war gases was associated with well-defined changes in the fluid and salt content of the blood and tissues without an apparent increase in the permeability of the blood vessels. Fluid and salt probably passed from the tissues to the blood in an attempt to com-

d This was done for a purpose having no connection with the present investigation and does not militate against the eonclusions drawn, inasmuch as this procedure did not noticeably change the development of edema.


pensate the latter for its loss in those constituents which mobilize in the lungs, resulting in edema. Later, if edema subsided, there might be reabsorption ot fluid and salt, a portion being redistributed to the tissues, the remainder being excreted through the kidneys. Such a hypothesis was supported by the correlation existing between the production and subsidence of edema, and the excretion of chlorides through the urine previously discussed.


From the fact that the lethal war gases exert a specific action upon the respiratory mechanism leading to impairment of the mechanism of respiration, it is obvious that distinct changes in the respiratory function of the blood are to be anticipated. This view is corroborated by the superficial observation that the blood changed in both its consistency and color. It became viscid and thick, and instead of possessing the normal bright red hue might assume a maroon color, often appearing almost black.

The subject of the respiratory function of the blood is of extreme importance, since upon its proper performance depends adequate tissue nutrition and continued existence of the organism as a whole. Oxygen starvation is an exceedingly serious condition resulting in impairment of all bodily functions, and if sufficiently grave, culminating in cessation of all vital activity.


By the oxygen capacity of blood is meant the cubic centimeters of oxygen in 100 cubic centimeters of blood which has been thoroughly aerated with atmospheric air at room temperature. Obviously the oxygen capacity of the blood is a measure of the hemoglobin present. From comparative studies of the total solids of the blood and oxygen capacity it appears that variations in oxygen capacity and hemoglobin can be accounted for by variations in the concentration of the blood by the lethal war gases.

With phosgene poisoning there were three distinct periods of fluctuation of the oxygen capacity. First, in all but a few dogs there was a diminution of the oxygen capacity immediately after gassing which lasted from four to seven hours. Secondly, there was an increase of the oxygen capacity which, in dogs that died, reached a maximum between eight and twelve hours, but in dogs that lived this value reached a maximum later. The third period marked a decrease in the oxygen capacity to normal which was reached at the twenty-fourth to thirtieth hour after gassing. The value for oxygen capacity sometimes fell to a figure slightly below normal during this later period. The above picture was the usual one to which by far the largest number of animals conformed. Occasionally, there was observed a case in which there was no diminution of oxygen capacity immediately after gassing, but instead a rapid increase to a maximum. Such animals usually died. Occasionally, also, an animal was observed whose oxygen capacity did not change at all, but fluctuated about the normal, during the entire period of observation. These animals usually lived. All of these oxygen capacity figures were paralleled by total solids figures so that it seemed justifiable to assume that one was not dealing with newly intruded corpuscles, but only with changes in the concentration of the blood. The oxygen capacity was determined in both venous and arterial blood, the values being the same in both cases.

e Then methods of Haldane and Barcroft were employed in the determination of changes in the gases of the blood.


The picture with reference to oxygen capacity was as follows: An immediate decrease, a subsequent increase to a maximum, followed by a return to normal or subnormal. These changes were independent of the concentration of gas to which the animal was exposed.

Immediately (up to one hour) after gassing with chloropicrin the oxygen capacity fell down markedly in 17 animals, slightly in 5, did not change in 2, rose slightly in 2, and markedly in 3. Treated statistically, this evidence indicates that there was a dilution of the blood in a majority of animals immediately after gassing with chloropicrin.

This first period of lowered oxygen capacity was brief (much shorter than the corresponding period in phosgene poisoning) and lasted less than two or three hours. Occasionally, this lowered oxygen capacity persisted in an animal for more than 24 hours. Two such animals in these experiments survived low concentrations of gas.

It is noteworthy that 12 of the 13 animals that died showed lowered oxygen capacity immediately after gassing. At low concentrations about one-half of the animals that showed this marked drop in oxygen capacity succumbed to gas poisoning. Of the eight deaths at high concentrations, seven showed diminished oxygen capacity immediately after gassing.

Following the initial short period of diminished oxygen capacity there was a quick rise above the normal figure. This reached its maximum between the twelfth and sixteenth hours in animals that survived. The maximum might come as early as the fourth hour and as late as the twenty-fourth. In dogs that died the maximum usually was found at the time of death. At high concentrations of gassing this maximum seemed to come rather early (4 to 10 hours). Finally, there was a third period in surviving animals when the oxygen capacity dropped to the normal or slightly below normal in 24 to 48 hours.

Briefly, then, the picture with reference to oxygen capacity was as follows: (1) An immediate decrease lasting only an hour or two; (2) a subsequent increase to maximum at death or 12 to 16 hours; (3) a slow decrease to normal or subnormal.

A study of the blood of dogs gassed with chlorine demonstrated that after gassing there was always a significant rise in the oxygen capacity.


By oxygen content is meant the cubic centimeters of oxygen in 100 cubic centimeters of blood just as it is drawn from the animal. The blood was drawn under oil to prevent contact with air, and in all the manipulation incident to the analysis contact with air was carefully avoided.

In the first period after gassing with phosgene in all dogs the oxygen content in arterial blood dropped slightly. In the second period the oxygen content tended to rise somewhat above normal in dogs that lived, while in dogs that died it rose slightly, then decreased steadily until death, when the value might be as low as one-half that of normal blood. In the third period in dogs that lived the oxygen content fell back to normal or slightly below.

The oxygen content may also be expressed in percentage of the capacity, which value is known as the percentage saturation. Expressed as such, the percentage saturation in the first period was within normal limits. In the second period the percentage saturation was often within normal limits at


first, but toward the last of this period and during the third period the percentage saturation would fall to a point slightly below normal in dogs that lived. In dogs that died the percentage saturation began to fall during the second period and at death was as low as one-half of the normal.

To summarize, the oxygen content of arterial blood taken as such did not vary greatly after gassing in dogs that lived. When taken together with the increase in oxygen capacity, however, it is apparent that the percentage saturation of arterial blood was reduced after gassing. In the first period after gassing with chloropicrin the oxygen content of the blood dropped somewhat in 19 dogs. It rose above normal in 6 dogs, and did not change in 4. This drop occurred in 90 percent of the dogs gassed at high concentrations, while at low concentrations only 60 percent showed this initial drop. In animals that died the arterial content of oxygen usually dropped as death approached. In the second period (i. e., after three or four hours from gassing) the arterial oxygen content tended to rise to a maximum, which appeared sometime between 8 and 24 hours. In the final period, the arterial oxygen content diminished, often to subnormal values.
In the period immediately after gassing the percentage saturation of oxygen was within normal limits. It rose or fell in a manner somewhat parallel to the oxygen content discussed above. In 12 animals the percentage saturation rose above the normal. In the period following the percentage saturation of the arterial blood usually dropped steadily for animals that died. In animals that survived the percentage saturation did not go below 70. The lowest figure was reached usually between 12 and 18 hours. After that it rose again to normal in 24 to 48 hours.

To summarize, the oxygen content of arterial blood taken as such did not vary enough in surviving gassed dogs to be appreciably significant. In dogs that succumbed the lowered arterial oxygen content was closely paralleled by the increasing weakness of the animal. The arterial oxygen saturation was actually increased immediately after gassing in a number of animals (12 out of 29)---mostly in those that died. No determinations were made with chlorine relative to the oxygen content of arterial blood.


In dogs that lived the value for oxygen content of venous blood after phosgene gassing dropped slightly immediately after gassing and thereafter fluctuated about a value which was below normal. In dogs that died, however, after the first decrease the value continued to drop rapidly until death, where the value for oxygen content in venous blood was often reduced to almost zero. Expressed as the percentage saturation, the value for venous blood was within normal limits during the first period, but dropped to a lower level in periods two and three in dogs that lived, while in dogs that died the percentage saturation fell during the second period to a value that sometimes was only one-fourth the normal.

The oxygen content of the venous blood after chloropicrin gassing dropped immediately after gassing in nearly all animals. In those that lived, the drop was on an average the same as in those that died. This average value was between 50 and 60 percent of the normal, regardless of the concentration of gas used. In the great majority of surviving animals, the venous oxygen content


did not return to normal in 48 hours. On the contrary, the value, though steadily increasing after the initial heavy drop due to gassing, remained usually at a low level for at least 48 hours. The percentage saturation of the venous blood presented essentially the same picture as described above for the venous oxygen content.

In a general way, then, the blood oxygen picture in chloropicrin poisoning was roughly similar to that found after phosgene gassing. If it is assumed that a drop in oxygen capacity is due to blood dilution and vice versa, it is found that a majority (22 out of 29) of the animals studied here showed blood dilution immediately after gassing with chloropicrin. The main difference in the initial dilution phases in phosgene and chloropicrin poisoning was in the duration of dilution. With chloropicrin it would last two or three hours, while with phosgene it was about six to eight hours long. The concentration of blood then followed in both types of poisoning. With chlorine gassing the oxygen content of venous blood usually showed a marked decrease which could be maintained for many hours.


From the data presented it is quite apparent that the changes in arterial blood must be intimately associated with alterations in blood concentration. In all three instances the change in oxygen capacity and arterial content closely approximated the corresponding fluctuations in blood concentration. With phosgene and chloropicrin the general character of changes under discussion was of the same kind, the difference being in the time relationships. The chlorine data differed in character from those of phosgene and chloropicrin in that with chlorine there was no initial drop in the oxygen capacity or arterial content. Instead there was an immediate progressive rise. A graph representing the general changes in phosgene is shown in Chart XVIII. The corresponding chloropicrin curve was so similar that it is omitted.

From the foregoing it may be concluded that changes in blood concentration adequately account for the observed alterations in oxygen capacity and arterial oxygen content. With phosgene and chloropicrin there was a corresponding initial fall synchronous with the dilution of the first period and a marked rise coincident with the increase in blood concentration. The rise in oxygen capacity and oxygen in arterial blood with chlorine corresponded with the changes in blood concentration.

When oxygen in the venous blood is considered, the changes observed with the three gases do not yield so simply to interpretation. It is true that the initial drop and the first rise in oxygen content corresponded with changes in blood concentration. The mechanism whereby this was brought about may involve several factors. Some of the factors which need consideration are edematous fluid in the lungs, circulation rate, concentration of the blood. It has been asserted that in the presence of lung edema a film of water forms over the pulmonary capillaries, through which oxygen must pass in addition to the capillary wall. This would result in the blood in the pulmonary veins being deficient in oxygen. The arterial blood, then, with an abnormally high value of oxygen capacity, does not have a corresponding high value for oxygen content, with the result that the percentage saturation drops in the arterial blood after the lung edema becomes well developed: that is, in the second period

and the first part of the third period. Other conditions remaining constant, such a state would result in the tissues being supplied with oxygen by blood subnormally saturated with oxygen and could lead to a drop in the content of venous oxygen.

This explanation, however, will not adequately account for the gradual continued fall in venous oxygen, nor do other conditions remain constant. The blood continues to concentrate to a point where its passage through capil- laries must become greatly impeded by the increased viscosity of the circulating fluid. In other words, the blood remains longer than usual in contact with the tissues and hence is robbed of an unusual quantity of oxygen. The continued increase in concentration ultimately reacts further upon the heart efficiency, less blood than usual being circulated in a given period, and even

CHART XVIII.- General changes in the oxygen capacity and content after phosgene gassing

though the oxygen content of this blood may be abnormally high there is finally an insufficient oxygen supply carried to the tissues. In other words, the concentration of the blood causes a circulatory failure which becomes progressively worse as blood concentration increases and the oxygen of venous blood becomes progressively low. Under these conditions the tissues must suffer from lack of oxygen and the nervous mechanisms tend to assume a condition of narcosis.   

The effect of oxygen want upon the heart will be to destroy its efficiency. the concentration and contracting force are decreased, and the heart may pass into a state of dilatation. As an accompaniment to the changes outlined


above, the blood pressure may fall markedly and the animal pass into a condition greatly resembling shock. The final analysis of the changes of oxygen in the blood leads back to the alterations of blood concentration as the primary cause.


The condition resembling shock exhibited by animals after chlorine poison- ing led to the study of acidosis. This was taken up by investigating the urine, the bicarbonate content of the blood, and the hydrogen ion concentration of the blood. The nature of the carbon dioxide-bicarbonate equilibrium in the plasma makes the sodium bicarbonate assume the r6le of a respiratory compound. Alkali or rather sodium ions are constantly being drawn from the tissue reservoirs to hold carbon dioxide in the blood and also constantly pass into the tissue reservoirs when the carbon dioxide tension in the blood is less. In other words, there is a considerable "alkaline reserve" in the animal body. An abnormal appearance of acid in the body leads to a reduction of this reserve. Hence the measurement of this sodium bicarbonate concentration in the blood gives an index to the reaction of the body; i. e., to the maintenance of the proper alkalinity or of an acidosis. The method of Van Slyke and Cullen was used. The blood was drawn without loss of carbon dioxide. The equilibrium <> 

was established at room temperature and at the tension of carbon dixoide in alveolar air.

The results with gassed dogs led to the conclusion that all animals gassed with chlorine showed an immediate acidosis (lowered alkali reserve) of more or less severity. This lowering of the bicarbonate content of the blood was an invariable result of gassing and though the degree of this acidosis was extremely variable, it carried no relationship to the concentration of chlorine to which the animal was exposed. Some interesting correlations between the ability of the animal to withstand the acute stage of chlorine poisoning and the bicarbonate value were observed. When the value fell from the normal, which ranges from 50 to 70, to below 40, the animal usually did not survive. Not all animals whose bicarbonate value rose after the first drop recovered; but all animals which survived the acute period had bicarbonate values rising sharply after the first drop. Again, in animals which were gassed more than once there seemed to be an overcompensation for the loss of alkali for each time the normal was higher than that before the previous gassing. This same sudden drop in the bicarbonate value was also observed when the concentration of the gas was very low. This loss of bicarbonate could be made good to the animal by intravenous injection or per os administration of sodium bicarbonate.

The causes of this acidosis are at best obscure. There are two theories each of which has its good points. There are certain facts which point to the absorption of chlorine as the cause for the acidosis. The fact that there was an immediate lowering of the bicarbonate value points to a cause which operated immediately and which was not cumulative. This appeared to be entirely independent of the concentration of the gas to which the animal was exposed. It was demonstrated that direct chlorine absorption probably did


not account for the observed changes in the tissue and blood. On the other hand, in the experiments on the bicarbonate value of the blood the amount of chlorine calculated to be necessary to produce the observed acidosis was exceedingly small--too small, in fact, to be determined by the most refined methods of analysis. Again, in the urine sometimes there was observed an immediate, though not large, increase in hydrogen ion concentration and usually in titratable acidity. These facts lead one to postulate a cause which is instantaneous in its action. The objection to the theory that the acidosis was due to the absorption of chlorine is found in the absolute lack of all correlation between the concentration of chlorine in the mixture breathed by the dog and the degree of acidosis produced. Since it was demonstrated that the musculature of the bronchioles contracted when the chlorine came in contact with it, this may account for the inability of the chlorine to penetrate into the alveoli and hence for the independence of the chlorine concentration and the loss of alkali.
The second theory is that of a carbon dioxide acidosis. When the chlorine struck the lung tissue more or less irritation, with the accompanying edema, resulted. This edema and the excessive secretion of mucus along the nasopharyngeal passages and trachea were always well developed at the end of the half hour gassing period. Along with this edema, as a result of the chlorine irritation, appeared the contraction of the muscles of the bronchioles. With the air passages contracted and the alveoli filling with edema fluid, the lung rapidly became seriously impaired for the purpose of allowing a free passage of oxygen into, and of carbon dioxide out of, the blood stream. This accumulation of the carbon dioxide in the blood, as the result of the inability of the carbon dioxide to leave the blood in the lung, if followed to its conclusion, would merely result in a readjustment of the ratio


by calling forth more alkali from the tissue reservoirs. This was exactly what happened after some five or six hours in dogs which had a fund of alkali to draw upon. In the meantime, however, there was a rapidly increasing tension of carbon dioxide in the blood without a compensating increase in alkali, and the result was a carbon dioxide acidosis. This condition would account both for the immediate moderate increase in the acidity in the urine after chlorine gassing, and for the high bicarbonate value in dogs which lived for 24 hours. Since the high bicarbonate level means merely a compensating mechanism and not the removal of the cause of the pathological condition, one can understand the reason for the fact that not all animals survived whose bicarbonate level returned to normal or above.

The objections to the above interpretation are that it has not been possible to demonstrate in any case an increased content of carbon dioxide in venous blood after gassing, while it may be shown that the ratio
was adjusted in considerably less time than the theory demands. Still another possible explanation of the acidosis is the insufficient oxidation which results


from aeration of the blood, the acid products of metabolism producing the characteristic acidosis. It is obvious, however, that the data at hand are insufficient to offer at present a final solution of the acidosis problem, but it is probable that acidosis will prove to be a resultant of the operation of all three factors discussed.


In addition to the determination of the acidity of the urine and the bicarbonate in the blood the hydrogen ion concentration of the blood is available as an indication of acidosis. The PH of blood is normally 7.4, which means that there is a slight preponderance of hydroxyl ions. The "buffer" value of blood is high and it is only in extreme acidosis that the reaction, as shown by the hydrogen ion concentration, changes, hence a study of this factor gives us valuable data on the degree of acidosis. The method of Levy, Rowntree, and Marriott was used. The whole blood was dialyzed against a neutral physiologic saline solution and the PH of the dialysate measured.

Whereas the PH of normal blood is 7.4, after gassing the value fell to 7.3 and 7.25. When the air was blown through the dialysate the value went up to 8.2 in all cases. Although the drop from 7.4 to 7.25 was small numerically, in view of the high "buffer" value of the blood and the well-known fact that it is the last tissue to change in its chemical characteristics, the observed values indicated a severe upset of the acid-base equilibrium. In the dialysate all of the salts in the blood were present, including sodium bicarbonate and ionized carbonic acid, so that there was the same ratio


in the dialysate as in the blood. If, after dialyzing, the carbon dioxide was blown out with air, the base formerly held as bicarbonate appeared as the more alkaline bicarbonate. If, however, the acidosis was due to fixed acids, the blowing out did not affect the hydrogen ion concentration; at least not to the extent observed. The fact that in every case the blowing out reduced the hydrogen ion concentration to the same value, 8.2, indicates strongly that the acidosis was caused by carbon dioxide rather than by fixed acids.

In brief, the blood picture after chlorine gassing was, on the basis of the data presented, as follows: The chlorine irritated the lung tissue, causing the bronchiolar musculature to contract and also edema to appear. As a result of the edema the blood became concentrated. The curtailed aeration, resulting from the edema, the concentrated blood, and the bronchiolar muscle contraction, resulted in a low degree of oxygenation of the blood; it resulted also in the inability to get rid of the carbon dioxide with the consequent accumulation which gave rise to a temporarily diminished alkaline reserve and to an increased hydrogen ion concentration. The decreased rate of circulation resulted indirectly in a very low carbon dioxide content of the venous blood. These altered conditions tended to return to normal in 24 hours.

The acid-base equilibrium in the blood was distinctly affected by gassing with phosgene, but in no definite direction. Whereas with chlorine there was invariably a drop in the bicarbonate value immediately after gassing, with phosgene the bicarbonate value dropped in some cases and increased in other


cases. In a normal animal the bicarbonate value is constant, never varying more than two or three volumes per cent. In dogs gassed with phosgene the variations above and below normal were large, sometimes as much as 10 volumes per cent. There was one definite tendency, however, and that was the drop in bicarbonate value as the animal approached death. This terminal acidosis was observed in all animals that died 10 or more hours after gassing. The PH value varied only slightly with the change in bicarbonate, the final drop as the animal approached death being the only pronounced change.

It may be concluded, then, that although phosgene caused a wide fluctua- tion in the bicarbonate value, there was no definite acidosis until the terminal stages. These appearances of acidosis must be referred, therefore, to the consequences of oxygen want in the terminal stages of phosgene poisoning and can not be regarded as a specific action of the gas. With cbloropicrin there seemed to be no marked eff ect of the gassing until some eight hours after gassing. During the first eight hours the values fluctuated about the normal. The PH determinations varied, in most cases, with the bicarbonate value. After the 8-hour period there was a gradual decline in both values, probably indicating an acidosis condition, though in neither case did the animal die within 24 hours. It appears, then, from the scant data on hand, that there was no immediate acidosis following poisoning by the lethal doses of chloropicrin.

It may be concluded, therefore, that the lethal gases fall into two groups with respect to the production of acidosis. With chlorine there was evidence of an immediate carbon dioxide acidosis which later may become readjusted, whereas with phosgene and chloropicrin, acidosis was apparent only in the terminal stages of poisoning and can hardly be accepted as being a specific response of the action of the gases. It is much more reasonable to regard this acidosis as a terminal acidosis induced by the condition, general depression, of the animal.


In the preceding pages an outline has been given of the changes that occurred in the organism as the result of exposure to the lethal gases. Restated briefly, the gassing had a definite influence upon respiration, pulse, temperature, blood concentration; water content of the lungs and tissues; chloride content of blood and tissues, with resulting changes in chloride excretion by way of the kidneys; red and white cells and hemoglobin of the blood: distinct alterations in oxygen of the blood, leading to dyspnea and partial asphyxia; the presence of acidosis at times, and a definite influence upon protein metabolism.

The effects of gassing as thus enumerated are so various and devious that an attempt toward correlation or the assignment of cause and effects seems at first glance well-nigh impossible. Further inspection of the data presented, however, brings to light one significant feature which stands out clear and distinct from all other effects induced by exposure to gas. This is the well-defined curve of changes in blood concentration. Upon the basis of alterations in blood concentration quite definite stages in gas poisoning may be outlined. These stages stand out most clearly with phosgene and, therefore, the picture presented by this gas will be considered first.



First stage.- In the first few hours (five to eight) after phosgene poisoning there was a notable decrease in the concentration of the blood. The decreased concentration occurred rapidly and then the blood gradually assumed the normal concentration. In this period there was sometimes a significant dilatation of the heart (observed by Eyster). Accompanying the decreased concentration of the blood there was a sharp drop in the chlorides of the blood and a marked increase in the chlorides and water content of the lungs. The chlorides of the urine increased immediately after gassing, reaching a maximum between the third and seventh hours, then decreasing. The heart beat was distinctly slowed at first, with a tendency to regain the normal or be somewhat above normal before this period was over. The immediate effect upon the respiration was a distinct increase in the rate. During this period the temperature showed a marked increase, attaining a maximum coincident with the termination of this period. Oxygen capacity, erythrocytes and hemoglobin followed a curve parallel with that of the changes in the concentration of the blood throughout all stages of phosgene poisoning. Oxygen content of both arterial and venous blood decreased significantly. The saturation of hemoglobin with oxygen decreased somewhat. In general, the decrease was more marked in the venous than in the arterial blood. In the first period an influence upon protein metabolism was not noticeable.

Second stage.- The period (five to eight hours) of blood dilution was followed by an interval during which the blood rapidly became concentrated to a point far beyond the normal value and remained near this level for several hours. In this stage the heart could be markedly decreased in size (Eyster). During the period of increased blood concentration the chlorides of the blood showed a tendency to regain the normal. The water and chlorine content of the lungs reached a maximum and then gradually decreased. The urinary chloride excretion was normal or subnormal. The heartbeat and respiration were both markedly accelerated.f The temperature, on the other hand, steadily decreased to a degree or more below normal. If the animal died in this stage the temperature might fall steadily up to the time of death. Most of the fatalities occurred in this stage. The oxygen content of arterial blood remained fairly stationary at a nearly normal value, whereas that of venous blood fell rapidly to a very low level. The saturation of hemoglobin with oxygen decreased rapidly in both arterial and venous blood, but the fall was greater in venous blood. There was no evidence of an influence upon protein metabolism.

Third stage.- After the period of increased concentration the blood gradually became more dilute until it was slightly under the normal value, which was eventually gained, and the animal recovered. The chlorides of the blood gradually regained the normal level. The chloride and water contents of the lungs followed a similar course. In animals reaching this stage the heart beat and respiration rose to a maximum and then gradually attained the normal. The temperature rose to normal or above in animals that eventually recovered. In animals that died during this period the heartbeat and respiration increased, but the temperature steadily fell. The oxygen content of

f In animals that were in a serious condition, although the rate of respiration was markedly increasing, there was a decrease in depth,  so that rapid shallow breathing existed.


arterial and venous blood tended to regain the normal. Chloride excretion by the kidney was markedly decreased, but later was much augmented. Coincident with the increased chloride excretion was a noticeable increase in the protein metabolism.

The interpretation which may be placed upon the different stages of phosgene poisoning is as follows: In the first stage there was a marked dilution of the blood. There are at least two ways in which this dilution may be explained. In the first place, it may mean an increased blood volume, the excess fluid finding its way into the blood from the tissues in response to the strong irritative stimulus exerted by the gas upon the respiratory tract. Or, secondly, a diluted blood would result if the red cells were removed in part and deposited in some organ or tissue. In these investigations no studies were made to determine actual changes in blood volume. Reports by Eyster and Meek,however, who made such estimations, tend to the conclusion that in the stage under discussion blood volume is not increased, and they account for the dilution of the blood on the hypothesis that red cells are stored in the lungs, at least temporarily. Whichever explanation is correct, it is certain that during the first stage two features may be quite prominent, namely, edema of the lungs and dilatation of the heart. Edema may be explained very readily on the hypothesis of increased blood volume, and it is possible also that such a condition might lead to a dilated heart. On the other hand, the deposition of corpuscles in the lungs by causing an obstruction in the circulation would lead to a dilated right heart. The relatively large transport of fluid to the lungs during this period, however, is not explained so easily by this hypothesis. Whichever hypothesis is accepted, edema of the lungs prevails, and there may be a dilated right heart.

In the second period edema has reached its maximum development, and here also blood concentration is at its height. The latter state is undoubtedly induced by the withdrawal of fluid which finds its way into the lungs. During the interval of blood concentration the blood volume is definitely decreased and the heart may be noticeably diminished in size (Eyster). This would pre- sumably result in a decreased efficiency of this organ and would lead to an inadequate circulation. Later, when the blood resumes its normal degree of concentration, normal heart action is reestablished.

The development of edema induces a mobilization of chlorides in the lungs at the expense of the chlorides of the blood, the lowered chloride content of which may also be explained in part by loss of chlorides through the kidneys, since at this period the output of chlorides in the urine is appreciably augmented. Later during the second period, the chlorides of the lungs reach a maximum, the blood content is not called upon and, therefore, an approximately normal blood chloride content mav be found which is maintained thereafter. This chloride retention by the lungs coincides with the fact that on the second day of phosgene poisoning the urinary excretion of chlorides is usually below normal. The period of readjustment now follows during which edema subsides in the lungs, and presumably both fluid and chlorides are demobilized by the lungs and find their wav into the blood. The excess of chlorides over the normal in the blood is eliminated through the kidneys, which would account for the large output on the third day after gassing.


The changes in oxygen capacity, erythrocytes, and hemoglobin followed the curve of alterations in blood concentration throughout the entire course of phosgene poisoning, which might well be anticipated. Oxygen content of arterial blood in general showed relatively unimportant changes, whereas that of venous blood progressively diminished throughout the first and second periods of phosgene poisoning. This may be explained in the first period by the fact of diluted blood and in the second period was undoubtedly caused by the longer contact of the blood with the tissues, induced by an inefficient circulation.

The respiratory changes were correlated with the impaired respiratory functions of the blood, such as lowered inhibition. The later rapid pulse was directly induced by the viscous character of the blood which caused oxygen want. Although specific data are lacking, it appears quite evident that there was distinct fall of blood pressure. One may assume a direct relationship between the heart's efficiency and temperature. Thus, in the first part of the first period the heart action was slow, there was inefficient circulation, and the temperature fell. Later, the greatly accelerated pulse was accompanied by a rise in temperature far above the normal. From this it would appear possible that the heart had temporarily overcompensated, resulting in an efficiency of the circulation above the normal level.

Now follows the period of concentration of the blood. This concentrated blood is, without doubt, more difficult to circulate through the body, and if the heart is doing only its normal work there will be, as a result of the thickened blood, a circulation of less than normal efficiency and such a condition apparently results in a falling temperature. In case the heart responds with a much higher rate during the period of concentration, so that even with the thickened blood it appears that a circulation of close to normal efficiency is being maintained, it will be found that the temperature is also well maintained.

In the animals which were less seriously affected and in which only a slight edema of the lungs developed, with a consequent slight loss of fluid from the blood, it was found that the temperature was well maintained provided tile heart rate was normal. However, even ill such cases the continuous, though slight, loss of fluid from the blood would eventually result in a concentration of the blood which would bring the circulation below normal efficiency, even with a high pulse rate, and the temperature would slowly drop until at about the twenty-fourth hour it was about 1oC., below normal. On the other hand, in the animals which were seriously affected, the blood concentrated very rapidly. The heart, even though the rate was maintained far above normal, was nevertheless not able apparently to maintain a circulation of normal efficiency, the temperature dropped very rapidly, and the animal died within less than 24 hours after gassing. In brief, thenr, it seems plausible that the temperature is directly related to the efficiency of the circulation and this in turn is determined, in part at least, by the concentration of the blood and the pulse rate.

This view appears to be further strengthened by the results obtained from the study of animals gassed with chlioropictin and chlorine. In both of these cases there was, in general, a state of concentration of the blood beginning immediately after gassing. Only in rare instances did a dilution of the blood occur and then it was only for a short time. From the first, then, in animals poisoned with these gases there obtained a condition in which the blood was


above normal in concentration and in correspondence with this the temperature remained below normal and the more seriously the animal was affected and the greater the concentration of the blood, the greater was the fall in temperature.

Phosgene poisoning has been considered in detail since it is unique in showing among its effects the initial period of blood dilution. At times chloropicrin presented a similar stage, but this interval was never so pronounced either in degree or length as obtained in phosgene poisoning. Usually a preliminary dilution period was lacking. It is this period that undoubtedly gives to phosgene the distinction of possessing a so-called "delayed action." Chlorine gas rarely, if ever, caused a period of blood dilution. In general, if one should consider the changes in blood concentration outlined for phosgene minus the initial dilution period, the remaining curve would represent fairly accurately the alterations occurring in the blood in both chlorine and chloroplerin poisoning. This would, of course, entail differences in time relationships, but under the conditions noted the changes in blood concentration of chlorine and chloropicrin would be accom- panied by the same general type of effects which are obtained with phosgene. Under these circumstances it appears superfluous to recite further the correla- tion of the effects of chlorine and chloropicrin poisoning.


It is generally assumed that death, in gas poisoning, is due directly to edema of the lungs, aided, of course, by the accompanying congestion. It has been said that death is caused by an individual literally drowning in the water of his lungs. The quantity of water present may reach as high a figure as a liter or more and such a conception of the cause of death seems quite obvious. On the other hand, one may well ponder whether death is usually induced in this way or whether there may be some other cause to which one may point with more certainty. The most obvious condition, other than edema, which could lead to death is the concentration of the blood. Of course, it is evident that edema and blood concentration are closely associated. Edema is assuredly the cause for blood concentration and thus indirectly, at least, brings about death, but it would appear that blood concentration is much more likely to produce death than is the presence of fluid in the lungs. There are, therefore, two possibilities open.

Death by edema could be caused by the prevention of an adequate oxygen exchange in the pulmonary blood. On the other hand, through extensive experiments of Winternitz,6 it is quite possible to introduce large quantities of fluid directly into the lungs of normal dogs without causing death, the fluid being absorbed with surprising rapidity. It must be conceded, however, that the conditions obtaining in the lungs of a normal dog and in those of a gassed animal are quite different, for in the experiments cited simple salt solution was introduced, whereas in an edematous lung the fluid more nearly represents blood plasma. Such a fluid would have a much greater tendency to inhibit adequate oxygen exchange than would a simple salt solution. The adherents of the idea that edema is the cause of death must ascribe death to asphyxiation. There is little doubt that well-developed edema does interfere with oxygen exchange of the pulmonary blood, but usually the efficiency of the arterial blood as an oxygen carrier is surprisingly high. It would seem a simple matter to put the question to the test experimentally. Thus, it might be assumed that if edema is the


cause of death, this operating by producing asphyxia, administration of oxygen should save the animal provided the oxygen could be absorbed. Such experiments were carried through in this investigation, and the results demonstrated that, even though the oxygen in the arterial blood may be raised and main- tained in the higher normal limits, death intervenes as usual. Then, again, some animals seemed to die with much less edema than others, and the different gases also possessed different degrees of ability in provoking edema. If edema is the cause of death it is difficult to explain why some animals, with an apparent excessive quantity of fluid in the lungs, should have survived. Death is caused by something more than simple inability of the blood to absorb oxygen, by something more than a physical obstacle in the lungs.

It seems quite logical to assume that blood concentration is immediately responsible for death. Blood concentration means a failing circulation, an inefficient oxygen carrier, oxygen starvation of the tissues, fall of temperature, and finally suspension of vital activities. The whole aim of treatment was to prevent blood concentration or else restore it to the normal level. When this was accomplished the animal survived in spite of the fact that the lungs might be very edematous. It may be stated, then, that in the presence of edema and a concentrated blood, entrance of oxygen into the circulation did not prevent death. On the other hand, restoring blood to the normal concentration enabled an animal to survive even though an extensive edema existed. Administration of oxygen under the last-named conditions undoubtedly made recovery easier. Therefore, while it is acceped that indirectly the edema of gas poisoning results in death, the immediate cause of death must be assigned to blood concentration.


(1) Mendel, L. B., and Rose, W. C.: Experimental Studies on Creatine and Creatinine. I. The Rôle of the Carbohydrates in Creatine-Creatinine Metabolism. Journal of Biological Chemistry, Baltimore, Md., 1911, x, No. 1, 213.
(2) Underhill, Frank P.: Studies in Creatine Metabolism. I. Possible Interrelations between Acidosis and Creatine Elimination. Journal of Biological Chemistry, Baltimore, Md., 1916, xxvii, No. 1, 127.
(3) Lamson, Paul D., and Roca, John: The Liver as a Blood Concentration Organ. Journal of Pharmacology and Experimental Therapeutics, Baltimore, 1921, xvii, No. 6, 481.
(4) Eyster, J. A. E., and Meek, W. J.: Experiments on the Pathological Physiology of Acute Phosgene Poisoning. American Journal of Physiology, Baltimore, Md., 1920, li, No. 2, 303.
(5) Bogert, L. J., Underhill, F. P., and Mendel, L. B.: The Regulation of the Blood Volume after Injections of Saline Solutions. American Journal of Physiology, Baltimore, Md., 1916, xli, No. 2, 189.
(6) Winternitz, M. C.: Collected Studies on the Pathology of War Gas Poisoning. New Haven, Conn., 1920, Yale University Press, 148.