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 XII




When an animal was exposed to the vapors of dichlorethylsulphide in high concentration it subsequently showed a complex of symptoms which maybe divided into two classes: (1) The systemic effects due to the absorption of the substance into the blood stream and its distribution to the various tissues of the body. These effects were not generally recognized. (2) The local effects on the eyes, skin, and respiratory tract. These were well recognized, and consisted mainly of conjunctivitis and superficial necrosis of the cornea: hyperemia, edema, and later, necrosis of the skin, leading to a skin lesion of great chronicity; and congestion and necrosis of the epithelial lining of the trachea and bronchi.

The most striking observation regarding the symptoms of dichlorethylsulphide poisoning was the latent period, which elapsed after exposure before any serious objective or subjective effects were noted. The development of the effects was then quite slow, unless very high super lethal doses had been inhaled.


The symptoms observed in dogs when subjected to the vapors of dichlorethylsulphide which might suggest absorption into the blood stream through lungs and skin, and a systemic effect, were vomiting, diarrhea, hyperexcitability and convulsions, and effects upon the heart. Moreover, the condition of the lungs and trachea found at autopsy of some animals dying from the inhalation of the gas was not sufficient to account for death.



In order to become familiar with the effects of the absorption of dichlorethylsulphide into the system, dogs were injected with the substance. The simplest method of introducing the substance was by subcutaneous or intramuscular injection of olive oil solutions. The effects observed on an unanesthetized animal from the injection of a lethal dose in olive oil were, after a latent period, salivation, hyperexcitability, and convulsions, diarrhea, slow and irregular heart, which became rapid before death, muscular weakness, and finally coma and death.
a The data in this chapter are based on the researches of the pharmacological research section, Medical Division, Chemical Warfare Service, conducted in the American University Experiment Station Laboratories, Washington, D. C., and published as follows: (1) Lynch, Vernon, Smith, H. W., and Marshall, E. K., Jr.: On Dichlorethylsulphide (Mustard Gas). I. The Systemic Effects and Mechanism of Action, Journal of Pharmacology and Experimental Therapeutics. Baltimore, Md., 1918, xii, No. 5, 265; (2) Marshall, E. K., Jr., Lynch, Vernon, and Smith, Homer W.: 11. Variations in Susceptibility of the Skin to Dichlorethylsulphide. Ibid, 291; (3) Smith, Homer W., Clowes, George H .A., and Marshall, E. K. Jr.: IV. The Mechanism of Absorption by the Skin. Ibid, 1919, xiii, No. 1, 1.


The following protocols illustrate the effects of subcutaneous and intramuscular injections in olive oil solution:

EXPERIMENT 1.- Dog V7, male; weight, 12.5 kilos
March 12.
10.15. Pulse 136, respiration 18.
10.28. Subcutaneous injection of 500 mgm. dichlorethylsulphide in 5 grams olive oil.
    10.45. Behavior normal. Licks point of injection.
11.00. Somewhat restless.
    12.00. Salivation, vomiting, diarrhea.
    1.15. Pulse 66, respiration 32.
    1.30. Muscular spasms, especially in hind legs. Convulsions. Animal staggers. Finally unable to walk.
    1.40. Intraperitoneal injection of 3 grams chloretone in olive oil.
    2.10. Pulse 66 (very strong), respiration 56. Still excitable.
    2.45. Pulse 64, respiration 36.
    4.15. Pulse 126, respiration 26. Given another injection of 500 mgm. sulphide.
    5.00. Animal conscious, but unable to stand.
    March 13. Found dead.
Autopsy.-Lungs and trachea appear normal; stomach contains bile and bloody fluid; intestines congested with few areas of hemorrhage and contain bloody fluid.
EXPERIMENT 2.-Dog V10, female; weight, 10 kilos.
9.50. Pulse 88, respiration 18.
    9.55. Subcutaneous injection of 1,000 mgm. of dichlorethylsulphide in 10 c.c. of olive oil, 5 c.c. on each side.
10.00. Pulse 90, respirations 20. Animal quiet.
    10.15. Restless.
    10.30. Pulse 118, respiration 44.
10.50. Pulse 112, respiration very rapid, slight salivation.
11.15. Salivation, rapid respiration, diarrhea.
12.00. Pulse 44, respiration 102. Animal has previously had very good control of muscles. Movements are now stiff, convulsive, and uncontrolled. Tries to struggle to feet, but topples over and paws ground convulsively.
1.00. Pulse 36 (very strong, but irregular), respiration 112.
1.45. Pulse 36, respiration 64. Very weak, unable to rise. Has vomited.
2.40. Pulse 54, respiration 82. Vomiting. Now perfectly quiet.
3.00. Pulse 72, respiration 66.
3.20. Pulse 108 (very weak), respiration 54.
3.30. Dies.
Autopsy.-Trachea and lungs appear slightly congested; stomach filled with bloody fluid; intestines congested and hemorrhagic; other organs appear normal.
EXPERIMENT 3.-Dog DM203; weight, 17 kilos.
September 23.
10.30. Pulse 88, respiration 18.
10.35. Intramuscular injection of 240 mgm. (14 mgm. per kilo) of dichlorethylsulphide in 12 c. c. of olive oil.
11.00. Behavior normal. Pulse 114, respiration 30.
11.30. No toxic symptoms as yet. Pulse 100, respiration 20.
12.00. Pulse 102, respiration panting; is becoming hyperexcitable; appears irritable and trembles slightly, but walks without ataxia.
12.45. Pulse 120, respiration 28. Walks uncertainly; salivated; pupils normal. Injected leg affected, seems sore, and dog refuses to use it in standing or walking.
1.10. Pulse 120, respiration 30. Injected leg no longer seems sore, hut is wholly useless.


1.45. Pulse 130, respiration 36. Though seemingly conscious, movements are decidedly convulsive. No diarrhea, great salivation or hyperexcitability. Pupils are now greatly dilated.
2.00. Pulse 114, respiration 29. Dog exhibits twitching and slow convulsive movements. Knee reflex in injected leg about normal; the leg is inactive during convulsive movements. Very irritable and snappy at times; then again is  very affectionate. After drinking heavily, pulse runs up to 150, but is very irregular.
2.45. Pulse 144, respiration 60. Shows slight salivation.
4.20. Pulse 174, respiration 66. Still shows slight convulsive tendencies.
September 24.
9.00. Pulse 192, respiration 48. Dog is conscious, but trembles greatly. No salivation or evidence of diarrhea. Pulse varies greatly during day.
  September 25.
1.00. No marked change in condition. Pulse running 150 to 180 throughout day, respiration 24 to 35.
September 26.
10.20. Pulse 156, respiration 32.
4.00. Dog found dead 77 hours after injection.
Autopsy.-Conjunctivitis; slight cutaneous edema at site of injection. Tracheal blood vessels congested; lungs apparently normal, with slight post-mortem changes. No hemorrhages in adrenal cortex. Liver, kidneys, and spleen normal; mucous membranes of gut hemorrhagic and bloody, chiefly in upper tract; stomach normal.
EXPERIMENT 5.-Dog DM202; weight, 11 kilos.
September 23.
10.30. Pulse 102, respiration 66.
10.40. Intramuscular injection of 220 mgnm. (20 mgm. per kilo) of dichlorethyl sulphide in 5.5 c. c. of olive oil.
11.00. Dog appears normal, except for panting. Pulse 138.
11.30. Pulse 100, respiration still panting.
12.00. Pulse 78, respiration 60. Marked salivation; appears spasmodically ataxic, with severe tremors; distinctly hyperexcitable.
12.45. Pulse 78, respiration very irregular. Salivation, but no diarrhea or vomiting. Exhibits severe convulsions. Pupils dilated. Injected leg stiff and does not enter into convulsive movements, which are of a clonic nature; between convulsive spasms, dog attempts to rise, but both rear legs are inactive.
1.10. Pulse 75, respiration 70. Passes small quantity of semiliquid stool without blood stain.
1.45. Pulse 68, respiration 36. No continued diarrhea.
2.00. Pulse 66, respiration 30. Heart sounds hammer-like, but very irregular in periodicity and intensity.
2.15. Pulse 90, respiration 28. Dog attempts to drink with great effort, but cannot rise or put nose in water. Continues to lap air. Portrays marked muscular weakness and incoordination.
2.30. Pulse 68, respiration 39. Eyelids twitching. Reflexes abnormally active. Heart irregular, missing every third beat.
2.45. Pulse 150, respiration 40.
3.00. Pulse 150, respiration 48. Heart sounds becoming very faint, with increased rate.
4.00. Pulse 162, respiration 38. Great muscular weakness. Almost unconscious. Seems to be passing into coma.
4.20. Pulse 192, respiration 48. Dog lies in coma.
September 24.
12.00. Found dead after 13 hours.
Autopsy.- Trachea slightly congested. Lungs show small circumscribed points of hemorrhage throughout; considerable post-mortem congestion; adrenals show slight cortical inflammation. Kidneys normal, except for slight postmortem congestion. External gut has too much post-mortem change to describe extent of external hemorrhage. Stomach is slightly congested. Very marked hemorrhage in the lumen of jejunum and duodenum; entire gut blood stained throughout. Injected leg showed great subcutaneous edema but no muscular changes beyond a slightly brighter color than in normal leg.


The intravenous injection was much more instructive but somewhat more difficult. The slight solubility of the substance in water (about 0.07 percent at 10° C.), and the rapidity with which an aqueous solution hydrolyzes were the main difficulties. This could not be overcome by the injection of an alcohol or acetone solution, for as soon as these solutions came in contact with water the sulphide was precipitated out as fairly large oil droplets. These difficulties were overcome by the injection of large amounts of a freshly prepared, cold, saturated aqueous solution.

The method of preparing the aqueous solution for intravenous injection was as follows: Five hundred cubic centimeters of 0.8 percent saline was cooled to 8 to 10° C. This was placed in a flask, about 1 c.c. of pure dichlorethylsulphide was added, the flask tightly stoppered, and shaken for about one minute. The contents of the flask were transferred to a separatory funnel and the oil allowed to settle. An oil film was present on the surface as well as globules at the bottom. About 400 c.c. of solution was removed from between the oil and film, care being taken to obtain a solution free from oil films or droplets. This was placed in a bath at 8 to 10° C. and used as soon as possible.
The solution prepared in this way contained about 0.7 mgm. per cubic centimeter.

The hydrolysis of the aqueous solution is a monomolecular reaction. At 10° C. only about 15 percent of the disulphide is hydrolyzed in 10 minutes, while at 37.5° C. over 97 percent is decomposed in the same time.
The symptoms elicited from an intravenous injection were similar to those observed from the subcutaneous injection of olive-oil solutions. While the solution was being injected, and for some time after, the animal showed no effects whatever. A record of the blood pressure, pulse, and respiration failed to show an appreciable effect at this time.

Within 10 to 20 minutes after injection, however, an increased salivation was noticed. This soon developed into a very free flow of rather mucinous saliva. The next symptom observed was usually a diarrhea, which might be accompanied ly vomiting. This diarrhea was present until the death of the animal, and a few hours after the injection the stools frequently contained blood.

After injection the respiration became rapid, and if anesthesia had not been used the animal showed a distinct hyperexcitability. At this stage he might be frightened by a slight movement of the hand or unexpected touch, and the eye reflex might be obtained by touching almost any part of the face. The gait soon became unsteady, movements of the muscles were spasmodic but uncontrolled, and apparently accompanied by tetanic contractions of the antagonist. The knee-jerk might be elicited by a touch of the finger. The animal soon became unable to walk or even to stand, and the violent spasmodic movements increased to the stage of convulsions, with extension of the hind legs and arching of the neck and back. The pulse, which might have been somewhat slowed, soon became irregular. Palpation suggested that the heart was dropping a beat occasionally. The dropped beat became more and more frequent until finally the heart was beating at one-half its former rate. If the chest was opened it was clearly seen that the ventricles were beating once to every two beats of the auricles. Later the rhythm might even become 1 to 3. Stimulation of the vagi showed some apparent hvperexcitabilitv. The blood pressure fell very slowly, and a few hours before death the heart resumed


its normal rate. At this stage, or before, it was found that the heart could not be slowed or in any way affected by strong stimulation of the vagi, although the respiration was easily inhibited. Section of the vagi had no effect upon the heart rate except a slight slowing. Apparently the vagus endings were paralyzed. The heart became feeble, there was a great dilatation in the splanchnic. area, a high venous pressure, and the arterial blood pressure was falling slowly. The convulsions ceased, and the animal lay in a coma. Death came quietly in less than 24 hours after the injection and was probably due to the weakening of the heart and the great dilatation of the vessels in the splanchnic area. Autopsy revealed a more or less intense congestion of the intestinal mucosa which might extend from the pylorus to the anus, and was frequently accompanied by hemorrhage into the lumen of the intestine. The condition suggested the excretion of the dichlorethylsulphide into the intestine. These effects upon the heart, the alimentary tract, and the central nervous system were quite characteristic and unmistakable.

The following protocols are representative of the course of an intravenous injection. In all cases the cold, saturated aqueous solution, prepared as described above, was used. The solution was usually injected within 10 to 15 minutes after its preparation. When injected such a solution contained about 0.5 to 0.6 mgm. of undecomposed dichlorethylsulphide to each cubic centimeter.

EXPERIMENT 6.- Dog, V9, female; weight, 8.2 kilos.
11.40. Pulse 88, respiration 16.
11.50-12.00. Injection into jugular vein of 120 c. c. fresh aqueous solution (14 c.c. per kilo).
12.05. Shivering. Pulse 160, respiration 66.
12.20. Slight salivation, has vomited. Pulse 70, respiration 74.
12.45. Diarrhea, no blood. Pulse 52 and irregular, respiration 108.
1.20. Highly excitable, convulsions, readily thrown into spasms-comlparatively quiet between.
1.45. Pulse 80, respiration 110.
2.15. Pulse 86, respiration 120.
2.45. Pulse 84, respiration 120.
3.45. Pulse 84, respiration 110. Struggling, proftuse salivatioin,has had diarrhea for some time.
3.50. Killed with ether.
Autopsy.-Trachea appears normal. Lungs very small area of congestion in one lobe, a few emphysematous patches. Esophagus and stomach appear normal. Duodenum, congested areas beginning at pylorus. Small and large intestine show distinct congestion.
EXPERIMENT 7.- Dog V17, female; weight, 5.5 kilos.
March 29.
3.10. Pulse 120, respiration 12.
3.12-3.20. Intravenous injection of 120 c.c. of cold, aqueous solution (22 c.c. per kilo).
3.25. Pulse 150, respiration 42.
3.35. Pulse 120, respiration 198. Trembling, good control of movements.
3.45. Vomiting; diarrhea.
4.05. Gait very unsteady.
4.30. Vomiting; fluid stools.
4.50. Struggling, convulsions begin; salivation.
5.00. Still in convulsions.
March 30.
9.00. Found dead.


Autopsy.-Trachea slightly congested; lungs appear normal; heart, large, endocardial hemorrhages; stomach contains bloody fluid; duodenum very deep red, bloody contents; small and large intestine deeply congested.
EXPERIMENT 8.- Dog V36, male; weight, 9.5 kilos.
April 22.
10.45. Pulse 90, respiration 11.
11.02.-11.05. Injection of 95 c.c. fresh, aqueous solution into jugular vein.
11.16. Slight salivation, vomiting.
11.25. More vomiting.
11.40. Hyperexcitability.
12.45. Convulsions.
1.00. Pulse 120, respiration 30, convulsions.

In the following experiments the dose given was too small to produce convulsions, but the other symptoms were observed.

EXPERIMENT 9.- Dog V20, male; weight, 11 kilos.
April 2.
12.18. Pulse 120, respiration 24.
12.18-12.23. Injection of 66 c.c. cold, fresh aqueous solution in jugular vein.
12.25. Pulse 116, respiration 21.
12.30. Perfectly normal.
1.05. Has vomited.
2.00. Pulse 120, respiration 20; normal movements.
3.00. Pulse 132, respiration 24; salivation; lying quietly, but leg muscle twitched a good deal.
April 3. Apparently normal; killed for autopsy.
Autopsy.-Lungs appear normal; duodenum congested and slightly hemorrhagic; small intestine contains bloody contents.
EXPERIMENT 10.- Dog V34, male; weight, 7.2 kilos.
   9.45. Pulse 96, respiration 12.
10.05. Intravenous injection of 36 c.c. cold, fresh aqueous solution (5 c.c. per kilo).
10.08. Pulse 108, respiration 14.
10.55. Has vomited.
11.05. Slight unsteadiness of gait.
2.25. Pulse 132, respiration 32. Gait normal.
EXPERIMENT 11.- Dog V19, female; weight, 17.3 kilos.
April 1.
12.55. Pulse 138, respiration 18.
1-1.05. Injection intravenously 34 c. C. of cold, fresh aqueous solution.
3.00. Pulse 78, respiration 12, quiet.
April 2.
    9.00. Depressed, very quiet, refuses food. Killed for autopsy.
Autopsy.-Lungs appear normal; duodenum congested, bloody mucous abundant; small intestine congested; large intestine normal.

The following protocol is illustrative of the symptoms of an animal injected very slowly. The symptoms were the same as when injection was made rapidly, but appeared to be somewhat more delayed and miller.

EXPERIMENT 12.- Dog V16, male; weight, 4.5 kilos.
1.30. Pulse 150, respiration 18.
1.55-2.55. Intravenous injection of 100 c. c. cold, fresh aqueous solution (22 c.c. per kilo).
2.54. Pulse 125, respiration 48, slight salivation.
    3.00. Pulse 120, respiration 78, salivation, seems excitable, defecated, had control of movements.
    3.30. Pulse 120, respiration 55, bad control of movements, diarrhea.
    3.45. Pulse 60, respiration 84, moves around fairly well.
    4.00 Movements Unsteady and spasmodic, convulsions begin.
    4.05. Unable to stand.
    5.00. Respiration 72, lying prone in convulsions.
    March 30. Found dead.


Autopsy.-Trachea normal; lungs, one lobe appears congested; duodenum, deeply blood-stained contents; intestines congested with blood-stained contents; heart, few small subendocardial hemorrhages.

The injection of saturated aqueous solutions, which have been allowed to come to room temperature or which have been standing for any time, showed much diminished effects, or, if several hours old, were entirely without effects.
This was tried by intravenous injection and also on the eye and skin. Such solutions were known to be partially or completely hydrolyzed into hydrochloric acid and dihydroxethylsulphide.


EXPERIMENT 13.-Dog V11; weight, 4 kilos.
March 22.
3-3.15. Injection intravenously of 160 c. c. aqueous solution 24 hours old (40 c. c. per kilo).
4.00. No effects.
6.00. No effects.
March 23.
9.00. Dog perfectly normal. No vomiting or diarrhea during night.
EXPERIMENT 14.-Dog EM970; weight, 8 kilos.
2.40. Pulse 130, respiration 18.
2.50. Intravenous injection of 10 c.c. per kilo of a saturated aqueous solution of dichlorethylsulphide which had stood for three and one-half hours.
3.00. Pulse 160, respiration 18.
3.50. Pulse 160, respiration 22; dog seems normal.
4.15. Pulse 164, respiration 20. Animal appeared perfectly normal while under observation for two or three days.
EXPERIMENT 15.-The cold, fresh, aqueous solution was applied to the forearms of five men. After standing at room temperature for two hours itwas again applied. Twenty-four hours later, four out of five of the men showed a distinct reaction from the fresh solution, none exhibited any effects from the solution two hours old.


Some of the effects which have been observed from the injection occurred when a dog was poisoned by inhalation of the vapor. When a dog was poisoned by inhalation of a very large dose of the vapor, practically all the effects obtained by injection (salivation, vomiting, bloody diarrhea, hyperexcitability and convulsions, slow, irregular pulse, becoming rapid before death, and paralysis of the vagi) were observed. This leaves little room for doubt that, in high concentrations, dichlorethylsulphide is absorbed through the lungs and produced its characteristic effects upon the body.

EXPERIMENT 16.-Dog V12, male; weight, 11.4 kilos.
March 21.
10.15. Pulse 168, respiration 36. Very active and playful.
10.30-11.30. In gassing chamber, exposed to 0.3 mgm. per liter of dichlorethylsulphide-restless, rapid respiration, excitement, salivation, vomiting
11.40. Hyperexcitable. Coughing.
11.50. Pulse 108, respiration 56.
1.00. Defecates. Marked muscular weakness, ataxia.
1.25. Convulsions.
1.45. Pulse 152, respiration 50.
3.30. Pulse 168, respiration 64; marked muscular weakness; salivation.
4.00. Vomits.
4.30. Pulse 176, respiration 52, pulse becoming feeble. Killed with ether.


Autopsy.-Mouth, esophagus, and stomach normal; duodenum shows congestion, contents bloody; small intestine has congested areas of mucosa; trachea, some slight membrane; lungs, one lobe congested and emphysematous; heart, kidneys, liver, and spleen, normal.

EXPERIMENT 17.-Dog V13, female; weight, 13.5 kilos.
10.00. Pulse 96, respiration 26.
10.05-11.05. In gassing chamber, exposed to 0.28 mgm. per liter of dichlorethylsulphide. Excitement, irritation of upper respiratory tract, salivation, and vomiting.
11.08. Loose stools.
11.35. Pulse 44, respiration 96; pulse irregular, expiration spasmodic.
12.35. Pulse 44, respiration 60; diarrhea, salivation, hyperexcitability with typical convulsions.
1.20. More violent convulsions.
  2.05. Pulse 68, respiration 28. Animal very weak.
2.30. Vomits voluminous, foamy fluid. Bloody diarrhea.
3.40. Pulse 70, respiration 30. Very weak, but still struggles some.
4.00. Killed with ether.
Autopsy.-Trachea slightly congested, contains some fluid; lungs, one lobe shows slight hemorrhage, few patches of emphysema; esophagus and stomach normal; upper intestinal tract deeply congested, congestion decreases until in ileum, normal; large intestine congested.


Furtherconvincing proof of the absorption of the sulphide through the lungs was furnished by the detection of one of the products of hydrolysis, dihydroxyethylsulphide, in the urine of animals poisoned by dichlorethylsulphide by inhalation. This hydrolytic product could also be found in the urine after the injection of mustard gas.

Dihydroxyethylsulphide was injected into a dog, the urine collected and examined for this substance. A positive test was obtained. Urine from a normal dog failed to give a positive reaction. This was necessary because of the occurrence of ethylsulphide and its precursors in normal dog's urine. This proves that the substance was excreted, in part at least, unchanged. The injection of dichlorethylsulphide was next tried, and the urine found to contain the dihvdroxyethylsulphide. The inhalation experiments were then performed.

EXPERIMENT 18.-Dog V47, female; weight, 13.2 kilos.
  May 19. Intravenous injection of 100 mgm. per kilo of dihydroxyethylsulphide in 10 c.c. saline.
May 20. No urine passed in cage; 230 c.c. urine by catheter. This was evaporated under diminished pressure to small volume, 25 c. c. of concentrated hydrochloric acid was added, and distillation carried out under diminished pressure. The distillate was extracted with ether, and the extract evaporated. The residue was tested by applying small amounts to the skin of several individuals. A typical mustard-gas reaction developed. This was considered evidence of the presence of dichlorethylsulphide.
EXPERIMENT 19.-Dog V49, female; weight, 21.4 kilos.
May 20-21. Over a period of 24 hours, injected subcutaneously with 60 mgm. per kilo of dichlorethylsulphide in olive oil, 10 mgm. per kilo at a time. Immediately after last injection, dog was catheterized and 200 c.c. urine obtained. This was concentrated as above, and tested for dichlorethylsulphide by applying to the skin of several men. No reaction obtained. It was then treated as above, and a slight amount of oily substance obtained on the distillate. The distillate
was extracted with ether, and the ether evaporated, leaving a globule of oil. This oil was placed in a small test tube, and the month of the tube held against the arm for 3 minutes. After 24 hours a distinct dichlorethylsulphide effect was obtained. The distillate contained mustard gas, and the urine dihydroxyethylsulphide.


May 24.
10.25-10.55. Gassed in continuous flow chamber for 30 minutes, 0.52 mgm. per liter.
May 25.
9.00. Pulse 102, respiration 34. Great muscular weakness. Unobserved, but seems to have been in convulsions.
10.00. Urine taken with catheter and examined for dihydroxyethylsulphide.
10.07. Dog dies in coma.

A severe reaction from the re-chlorinated product was not obtained, but a sufficient reaction was obtained to furnish evidence of dichlorethylsulphide being present.

It was found possible by the application of 0.1 to 0.2 gram per kilo to the skin of dogs to obtain the characteristic effects of the absorption of mustard gas; salivation; vomiting; diarrhea; hyperexcitability; rapid, feeble pulse and depression, but neither convulsions nor slowing of the heart. The product of hydrolysis of mustard gas was also detected in the urine.

EXPERIMENT 21.-Dog V90; weight, 16 kilos.
11.00. Normal pulse 108, respiration 36. Chest and abdomen shaved; 1.6 grams of pure dichlorethylsulphide rubbed into skin with glass rod. Animal placed so that draft prevented absorption by inhalation of the vapors.
11.30. Second application; 1.6 grams rubbed in (total quantity, about 200 mgm. per kilo).
12.00. Pulse 96, respiration 150.
1.40. Pulse 138, respiration panting.
5.00. Pulse 168, respiration panting. Dog placed in clean metabolism cage. June 4.
9.00. 200 c. c. of urine collected overnight; examined for dihydroxyethylsulphide. Positive findings. Pulse 198, respiration 36. Dog depressed; slight salivation.
June 5.
9.00. Pulse 186, respiration 48. Dog etherized.



The latent period in the development of the effects of dichlorethylsulphide, either local upon the eyes, skin, or respiratory tract, or systemic upon the heart, nervous system, and digestive tracts, suggests that the substance may be altered in the body before exhibiting its characteristic actions. In fact, the absence of any immediate effects when the aqueous solution was injected directly into the blood stream makes this assumption almost imperative. The simplest chemical change which this substance undergoes in vitro is hydrolysis into hydrochloric and hydroxyethylsulphide. That this change takes place in the animal organism was shown by the detection of dihydroxyethylsulphide in the urine. However, an injection of a hvdrolyzed solution of dichlorethvlsulphide was without effect. The dihydroxyethylsulphide, when applied pure to skin of man and dogs, produced no irritation whatever.c As much as 0.3 gram per kilo was injected intravenously into a dog without producing any apparent effect, immediate or remote. An injection of 1,400 mgm. per kilo caused only a slight stupor and loss of coordination, with a quick return to normal, and none of the symptoms of the dichlorethylsulphide were present.

c Victor Meyer states that this substance is nontoxic. Berichte der deutschen chemischer Gesellshaft, Berlin, 1886, xix, No. 3, 3259.


EXPERIMENT 22.- Dog EM575, male; weight, 7.8 kilos.
November 18.
2.30. Normal pulse 80, respiration 20.
2.45. Intravenous injection of 1,400 mgm. per kilo of dihydroxyethylsulphide in 100 c. c. of water (total, 10,920 grams). Injection followed by salivation and slight nausea.
2.50. Dog shows slight ataxia. Pulse 120, respiration 20.
3.10. Pulse 110, respiration 20.
3.50. Pulse 110, respiration 20. Dog appears stupid, and exhibits as light ataxia similar to light alcohol poisoning. No serious symptoms apparent.
5.00. Pulse 105, respiration 20. Dog quiet, perhaps slightly depressed. Eats and drinks with indifference.
November 19.
9.00. Dog normal. Pulse 98, respiration 24.
November 20.
9.00. Dog normal. Pulse 105, respiration 30.
EXPERIMENT 23.-Dog EM576; weight, 5.9 kilos.
November 19.
10.30. Normal pulse 100, respiration 20.
10.45. Intravenous injection of 200 mgm. per kilo (total, 1, 200 mgm.).
Dihydroxyethylsulphide in 50 c. c. of water. Animal exhibited no symptoms whatever. Observed for three days.
EXPERIMENT 24.-The pure dihydroxyethylsulphide was rubbed into the skin on four dogs and five men and produced no irritation, either immediate or remote.

This product of hydrolysis is not responsible for the effects of mustard gas. The other product of hydrolysis is hydrochloric acid and is not a very toxic substance. Relatively large amounts can be injected intravenously without producing any marked effect. This is readily understood. When injected intravenously it is immediately neutralized by the "buffer" action of the blood. The blood does not become acid and the tissues are never really exposed to the acid. When strong solutions are placed on the skin or mucous surfaces, or injected into the tissues, an irritating effect is noticed. Hydrochloric acid, however, injected in very large doses, does produce very definite effects upon the animal and can cause death. Both products of hydrolysis of mustard gas are very readily soluble in water and very sparingly soluble in organic solvents, or, in other words, have a low lipoid solubility or partition coefficient. It would be expected from this that they would not readily penetrate cells. Harvey has shown this to be true for hydrochloric acid.1


Dichlorethylsulphide is very slightly soluble in water and very freely soluble in organic solvents, or has a high lipoid solubility or partition coefficient.d It would therefore be expected to penetrate cells very readily. Its rapid powers of penetration are practically proven by its effects upon the skin. Having penetrated within the living cell, it would undoubtedly hydrolyze. The liberation of free hydrochloric acid within the cell would produce serious effects and might account for the actions of dichlorethylsulphide. The mechanism of the action of dichlorethylsulphide may be summarized as follows: 1. Rapid penetration of the substance into the cell by virtue of its high lipoid solubility. 2. Hydrolysis by the water within the cell, to form hydrochloric acid and dihydroxyethylsulphide. 3. The destructive effect of hydrochloric acid upon some part or mechanism of the cell.
d Attempts to determine accurately the partition coefficient of this substance are unsuccessful, due to the rapidity with which it hydrolyzes. It appears to be over 200, using xylene and water at 20 C.


Although hydrochloric acid does not penetrate cells readily and is easily neutralized by the buffer action of the fluids of the body, one might expect by flooding the body with large quantities of acid to produce some of the characteristic effects of mustard gas. Stimulation of the respiratory center is a well-known effect of acid. Convulsions and salivation may be produced by injection of hydrochloric acid and it was found possible to produce slowing of the heart by rapid injection of this acid.

The delayed action of mustard gas may be explained by the formation of some compound with some constituent of the blood. However, blood taken from dogs which had been poisoned with mustard gas and were exhibiting typical symptoms at the time injected into normal dogs, produced no effect. Serum treated in vitro with mustard gas and allowed to stand and then injected into a dog produced no effect. The fluid which is formed in the vesicles and blebs produced by the application of mustard gas to the skin produces no mustard-gas effects.


By the administration of large quantities of sodium bicarbonate, both intravenously and by mouth, the symptoms following an intravenous injection of dichlorethylsulphide could be delayed, and convulsions, but not death, were prevented.

Ten cubic centimeters per kilo of the fresh, aqueous mustard-gas solution given intravenously always caused convulsions, violent other symptoms, and death. A series of 11 dogs injected with this dose of the solution, were treated with 10 c.c. per kilo of 5 per cent sodium bicarbonate by mouth and intravenously every hour for five or six hours. In one of these animals no benefit was obtained; in 6, symptoms were delayed and milder; in 4, convulsions were prevented. Death was never prevented, although sometimes apparently delayed. The action of mustard gas on the heart appeared to be increased, and most of the animals showed a very slow, failing heart before death. Since sodium bicarbonate (Harvey) is known to penetrate cells only with difficulty, much benefit was not expected. Numerous amines, especially those of high lipoid solubility, have been tried, but thus far the effects have not been consistent. A substance possessing the same physical properties as mustard gas, but slowly yielding alkali on hydrolysis, would be ideal to try for treatment.


The fact that the velocity of hydrolysis of dichlorethylsulphide is very much decreased by lowering the temperature, suggested trying the effects on animals at a high and low temperature. The influence of temperature was quite marked on drugs which underwent a change in the body before acting. The lethal dose of atoxyl and colchicine was increased markedly for frogs when the temperature was lowered, and decreased when it was raised to 37° C.2 Atoxyl is supposed to be reduced in the body before acting and colchicine is oxidized. According to the theory which has been advanced for the mode of action, one might expect that at a low temperature the rate of liberation of hydrochloric acid in the cell might he slow enough to be nontoxic, for a dose which would prove fatal at a higher temperature.


Fish were used for the experiments. Those kept at a low temperature survived the same dosage which proved fatal to those kept at room temperature.

EXPERIMENT 25.-Seventeen healthy catfish were exposed, in groups of four to six, to dichlorethylsulphide by immersing in a quarter saturated solution of mustard gas in tap water at 100 C. for five minutes. Ten of these were placed in water at room temperature (24° to 26° C.), and ten died (100 percent) within 28 hours. The other seven fish were kept at 8° to 10° C., and one died in 96 and another in 108 hours, while the other five survived for 5 days, when observations were discontinued.

EXPERIMENT 26.-Fifteen goldfish were exposed, as in experiment 25, to a half-saturated solution for 10 minutes. Eight were transferred to a bath at room temperature (25° C.) and seven to a bath at 8° to 10°C. Of the first group, four died in 26 hours, two died in 56 hours, and two died in 6 days. Of the second group, all survived 17 days, when observations were discontinued.

It is quite noticeable that in the catfish which were kept at room temperature hemorrhages generally developed in the fins and tails and in the ventral surface of the body after 16 to 20 hours,e while in those kept cold, no hemorrhage was ever observed. Symptoms, which were marked in the first group, were never observed in the second.

Although this evidence fits in perfectly with our theory, it is well known that the effect of drugs is changed by decreasing temperatures.The following experiments indicated that hydrochloric acid was toxic for fish whether the fish were cooled after exposure or kept at room temperature.

EXPERIMENT 27.- Eight catfish were exposed to 0.05 percent hydrochloric acid for five minutes. Four were placed at room temperature (25°C.), and four placed at 8° to 10°C. Of the first group, all died within 40 minutes; of the second, three died within 30 minutes, and one in 150 minutes.

EXPERIMENT 28.-Sixteen goldfish were exposed in the same manner as in experiment 27. Eight were kept at room temperature (20°C.) and eight at 8° to 10°C. Of the first group two died in 1 hour, and six in 1 hour and 20 minutes. Of the second group all died within 1 hour.

An experiment on atropine and one on sodium cyanide indicated that these substances were just as toxic for fish whether they were kept at ordinary te npermiture or cooled, after exposure.

EXPERIMENT 29.-Four goldfish were exposed for 10 minutes to 0.01 percent atropine sulphate solution. Two were placed in a bath at 8° to 10°C. In the first group, both died in 6 to 7 hours; in the second group, both died in 5 to 7 hours.

EXPERIMENT 30.- Six goldfish were exposed for 10 minutes to 0.48 percent solution of sodium cyanide. Three were kept at 20° C. and three at 8° to 10°C. They died in 3, 5, and 8 hours, and 4, 6, and 8 hours, respectively .

Catfish survived fifteen times the concentration of hydrochloric acid present at the end of 5 minutes in the half-saturated water solution of mustard gas, and three times the concentration present at the end of 24 hours. Fish survived 15 minutes' exposure in this solution after it had stood 1 hour at room temperature.

It is evident that these experiments on fish tend to substantiate the theory of intracellular liberation of acid.
e It is interesting to note that dichlorethylsulphide appears to act on the skin of catfish and not that of goldfish.
f Sollman (A Manual of Pharmacology. 1917, 94) holds that digitalis, veratriin, nicotine, strychnine, tetanus toxin, chloral, and alcohol are rendered more active, morphine and curare less active, by raising the temperature for cold-blooded animals.



Various compounds related to mustard gas were prepared at one time or another by the offense chemical section of the Chemical Warfare Service. A cursory survey of their lipoid solubility, rates of hydrolysis, and pharmacological effects was made.

The lipoid solubility was estimated by using xylene and water. Victor Meyer3 noted that ethylsulphide was inactive, while the B-monochlorethylsulphide was less active than mustard gas. Monochlorethylsulphide and the two isomeric B-B-dichlorpropylsulphides were all highly (though in different degree) lipoid soluble and all hydrolyzed more or less rapidly in aqueous solution. All were active skin irritants, and more or less toxic on inhalation, producing lesions and symptoms comparable to those of mustard gas.

The theory of the intracellular liberation of hydrochloric acid as the mechanism of action of dichlorethylsulphide, explains all the experimental facts thus far observed. The histological changes in the skin have been stated to resemble hydrochloric acid burns. According to Warthin and Weller: 4 "The lesion is a chemical burn unlike that produced by heat, electricity, or the ordinary corrosives such as sulphuric, nitric, and hydrochloric acids, or strong alkalies. Of all these agents, the effects are most closely allied to those of hydrochloric acid, but are much greater in intensity." Lillie, Clowes, and Chambers,5 in a study of the action of mustard gas on marine organisms, especially starfish eggs, have obtained evidence which supports our interpretation of the action of dichlorethylsulphide, e. g., an intracellular liberation of hydrochloric acid as the toxic factor.

A theory to explain the edema, fat, infiltration, multiple hemorrhages, and necrosis of the central portion of the liver lobule " by the '"severe tissue effects by the halogen acids formed in the tissues" has been advanced by Graham 6 for the delayed poisoning by chloroform and other alkyl halides. Graham's evidence is briefly as follows: Similar morphological changes can be produced especially in the liver by injection of hydrochloric acid into the portal vein. Sections of the chloroformed liver show an acid reaction with neutral red or Nile blue immediately, whereas such changes occur more slowly in the liver of normal animals after death. Carbon tetrachloride is more effective than chloroform and methylene chloride less effective in producing liver necrosis; this is in the same order as the amount of hydrochloric acid that can be liberated. Simultaneous injection of sodium carbonate with the chloroform anesthesia prevents or decreases the liver injury and other effects. Other alkyl halides give a typical morphological picture of chloroform poisoning. That these substances form halogen acids is shown by the excretion of the neutral salts in the urine. Chloral, which does not yield an appreciable quantity of hydrochloric acid in metabolism, does not produce these changes. While Graham did not discuss the question of how the acid is formed, he suggested Nef's 7 ideas of dissociation, forming bivalent carbon.

It is evident that the theory suggested in this chapter is somewhat different. Graham 6 did not consider intracellular liberation of acid, and was dealing with substances which do not readily yield hydrochloric aci(l on contact with water, but must undergo relatively slower metabolic changes by the tissues. The effects of mustard gas are not similar to those of delayed chloroform poisoning and, moreover, chloroform is a substance of an entirely different order of toxicity.


It is not impossible to reconcile these two different types of poisoning as both being due to hydrochloric acid. Graham 's suggestion that "the liver is an organ which most strikingly manifests the chloroform necrosis, and that this organ is also the site of a most active metabolism, harmonize well with this hydrochloric acid theory," may explain the difference. As stated above, chloroform must be slowly metabolized to produce hydrochloric acid. By oxidation chloroform yields phosgene and hydrochloric acid. Phosgene readily yields hydrochloric acid on hydrolysis.

One is inclined to agree with Graham's statement that the halogen acids are suggested to be important factors rather than the only factor involved. Other acids must play a part, and possibly other substances than acids are involved.6

With mustard gas and analogous compounds where the radical, other than the acid radical, is nontoxic, the entire responsibility can be laid to the intracellular production of acid. The question of cell penetration, as well as ease of hydrolysis and lipoid solubility, has to be considered in dealing with a series of compounds.7

It is interesting to note that most of the war gases can readily yield a halogen acid by hydrolysis. Whether the intracellular liberation of acid will explain their relative toxicity must be decided by future work. Their partition coefficients, their rate of hydrolysis, their volatility and other physical chemical properties may explain the differences in localization and intensity of action. The facts which can be gathered from the literature concerning dimethylsulphate, seem to indicate that its toxic action may be due to the intracellular liberation of sulphuric acid. The symptoms recorded in the literature due to this poison are very suggestive of mustard-gas poisoning-local redness and edema, generalized toxic and clonic convulsions, coma and death.8 The effect on the eyes9 and the clinical descriptions of several factory cases of poisoning resemble mustard gas 10. Both Weber 8 and Michiels,8 who worked experimentally on animals with this substance, discuss the possibility of the effects being due to acid poisoning. They decide against it because the amount of acid liberated is too small, the blood gases of the poisoned animal were found normal, sodium carbonate did not help poisoned animals, and it is doubtful whether time elapsed for a toxic dose to hydrolyze. Michiels found, moreover, that serum treated with a toxic dose of dimethylsulphide, incubated and injected into a rabbit, was nontoxic. So he agreed with Weber that the molecule is toxic and not a decomposition product. The evidence appears to be very suggestive for an intracellular liberation of sulphuric acid as the toxic factor. It was found that the partition coefficient is about 1.9, using xylene and water at 20°C., and that this substance hydrolyzes much more slowly than dichlorethylsulphide. It would be expected that the toxicity on intravenous injection would be less than mustard gas. Preliminary experiments indicate that 50 mgm. per kilo are necessary to produce the same symptoms as are given by 6 mgm. per kilo of dichlorethylsulphide. The symptoms appeared much more quickly in the case of the dimethylsulphate. This would indicate that mustard gas is prevented for a time from hydrolysis by being held in lipoids, while dimetlhylsulphate, having a much lower partition coefficient, is more readily transferred to an aqueous phase for hydrolysis. Certain facts noted later as having been observed with respect to the action of dichlorethylsulphide on the skin also led to this conclusion.



The results of the experimental work which has been considered may be summarized as follows:

1. Dichlorethylsulphide is absorbed throughthelungs and produces definite, characteristic, systemic effects.
2. The symptoms of injection of the substance are salivation, vomiting and diarrhea, tonic and clonic convulsions, slow and irregular heart, followed by a rapid pulse, and stimulation of the respiration.
3. A dose of six milligrams or less per kilo, injected intravenously in aqueous solution, proves fatal for dogs.
4. Dichlorethylsulphide appears to be excreted in the urine, in part at least, as dihydroxyethylsulphide, which has been shown to be a comparatively nontoxic body.
5. The lesions in the intestine suggest that excretion of the substance may also take place here.
6. The dichlorethylsulphide penetrates the cells, and in the aqueous phase of the cell, hydrolyzes to hydrochloric acid which is responsible for the damage.
7. Sodium bicarbonate in large doses somewhat alleviates the symptoms, but does not prevent death.
8. Fish are much less susceptible to this substance when kept at a low temperature after exposure rather than at room temperature. The hydrolysis of the substance is much slower at a low temperature.
9. Monochlorethylsulphide, and dlichlorpropylsulphide are lipoid soluble and easily hydrolyzed. Both act similarly to mustard gas.
10. The evidence leads to the view that dimethylsulphate acts by intracellular liberation of sulphuric acid.



Every investigator who has worked with d1ichlorethylsulphidle has noticed that some individuals are much more susceptible to the substance than others.Victor Meyer, its discoverer, noted that his laboratory assistant was greatly affected by exposure to this agent, while he himself was not.11 Similar observations have been made by many workers in laboratories. Despite the great differences in susceptibility, it is very improbable that any man or beast is immune to the action of mustard gas. It was possible to demonstrate beyond a doubt the individual variation in susceptibility, and methods were evolved by which determinations of susceptibility could be made.


Two methods were used for determining the cutaneous sensitivity of individuals to dichlorethylsulphide. The first of these methods consisted in exposing skin to the vapors of the substance under constant conditions and determining the minimum time of exposure which was necessary to produce a visible reaction within 24 hours. The apparatus which was used in this method consisted of a small test tube (1 cm. by 10 cm.) containing a cotton plug saturated with pure dlichlorethysulphide held by means of a rubber stopper in a larger test tube which was filled with water. When not in use the smaller


tube was closed with a cork stopper. On being first prepared the tubes were allowed to stand unstoppered for about 24 hours in order to remove any volatile impurities that might be present. Before use the whole apparatus was placed in a constant temperature water-bath. In all this work 20° C. was taken as a constant-temperature for making exposures. The skin was exposed to the vapors in the smaller tube by holding the mouth of the tube firmly against the skin. Exposures were made for different lengths of time in order to determine the shortest exposure which would just produce a visible reaction, the minimum burning time. The reaction was not visible for some hours, and was noted 24 hours later. The reaction consisted in a circular area of erythema which was usually uniform. In certain cases a group of small, red spots might be all that was noticed. Longer exposure than the minimal gave rise to edema as well as erythema, and a still longer one to subsequent small vesicles or a well-defined blister over the entire area exposed. In this method the skin was exposed to supposedly saturated vapor at 20°C.

The second method which was used for studying the sensitivity of individuals consisted in applying to the skin standard solutions of mustard gas in paraffin oil. For most purposes the best solutions were a 1 percent (1 gram of pure dichlorethylsulphide in 100 c.c. of oil), a 0. 1 percent and 0. 01 percent. A small drop of each of these solutions was applied to the skin of the forearm and the arm allowed to remain uncovered for about 10 minutes. The presence or absence of a positive reaction was indicated by the appearance or absence of erythema 24 hours later.

Within certain limits the amount of the solution of a given concentration of mustard gas which was applied to the forearm made no difference in determining whether the reaction was positive or negative. It made a difference in the intensity of the burn, a large amount of 1 percent solution causing a blister, while one-tenth the quantity did not. For this reason it was advisable not to employ too large a quantity of the stronger solutions in making the test. The following experiments were made, applying at the same time the amount of solution of mustard gas in paraffin oil left on a glass rod after dipping it in the solution and shaking it off, and then applying about ten times the quantity. The reaction from the larger quantity is given in brackets. (Tables 28 and 29.)

  TABLE 28.- Irritating effects of dichlorethylsulphide applied locally


 TABLE 29.- Irritating effects of dichlorethylsulphide applied locally

The question has been raised as to what length of exposure was given to the oil solution on the arm. Within 5 minutes, where only al small amount of solution was left on, absorption or evaporation was complete. When a larger amount of solution was applied, rubbing off the excess 5 minutes after application made no difference in whether the reaction was positive or negative. When the test was applied the subject left the sleeve rolled up for about 10 minutes and then continued his usual duty.

Twenty-four hours after making the test was a convenient time for reading the results. The following figures show that any time between 24 to 48 gave the same result. Less than 24 hours' readings should not be taken as some individuals were rather slow in reacting.

Paraffin oil was chosen as the solvent for making the test, for the following reasons: It was not very volatile; the solution of mustard gas in it was fairly stable; it did not spread greatly when applied to the skin; and it was probably not readily absorbed. Mustard-gas vapor was given off, however, from a 1 percent solution in paraffin oil. A small amount of this solution placed in a test-tube and held against the skin of a sensitive individual (for 10 minutes) caused erythema. Solutions of mustard gas in other solvents than paraffin oil were trieo on a selective number of individuals of the resistant, average, and sensitive types. The solvents used were linseed oil, cottonseed oil, kerosene, absolute alcohol, and paraffin oil as a control. The solution in absolute alcohol was the more reactive, while those in linseed, cottonseed, and kerosene oils were less reactive than that in paraffin oil. The greater reactivity in alcohol was probably due to its volatility leaving a stronger solution on the skin a few seconds after it was applied.

The skin of different parts of the body probably reacted slightly differently from mustard gas, although this was not tested fully. The solutions were always placed on the inner surface of the forearm. One point which is to be noted is that areas of skin in the neighborhood of old mustard-gas burns were more sensitive than other areas. They should be avoided in making the test.

The first method undoubtedly gave the most accurate results, although the second was very much simpler and more rapid in its application. Numerous


tests were made on individuals by these methods in order to determine if each method gave the same result. Men who were sensitive to one test were sensitive to the other, and vice versa. Roughly, it may be said that a man who gave a reaction to the 0.01 percent solution would show a reaction from a 15-second (or less) vapor exposure, while a man who showed no reaction to the 1 percent solution would be negative or only faintly positive to a 4-minute vapor exposure.

A method such as the second method described above might be applicable in the field or factory, being used to detect individuals who are especially sensitive to mustard gas so that they may be removed from areas which are exposed to this gas. It would be an easy matter to examine thousands of men and determine quickly which ones were especially susceptible to the gas and which were less susceptible. A sensitive individual living in a region where mustard gas is present is practically certain to be a casualty. It would be far easier to remove him before injury. This idea may sound very impractical, but it does not seem less practical than applying ointments to the body for protection.


This method was applied to only a few individuals working around the laboratory. The results are shown in Table 30.

TABLE 30.- Time of exposure to vapor tests required to produce a visible reaction

It is interesting to note that the very extreme cases had a minimum burning time of 1 second and 10 minutes, respectively. In other words it may be stated roughly that one of these individuals was six hundred times as sensitive as the other. The man who gave a reaction from a 1-second exposure blistered from a 5-second exposure. or one to which the vast majority of individuals will give no reaction. When first found this subject blistered from the application of the 0.01 percent solution in paraffin oil. Fortunately, it was possible to obtain a complete history and physical examination, but nothing of note was found. This man had never been exposed to mustard gas nor had he received even a small experimental burn when first tested.


As has been pointed out, the oil-solution method of determining sensitivity was very much easier and more rapid in its application than the vapor-test


method. Results were obtained on rather large groups of individuals, using this method. Table 31 summarizes the results obtained on a group of men at the American University Experiment Station. Practically all these men were indoor workers, and a great many of them had been more or less exposed to mustard gas.

TABLE 31.- Skin irritation from local application of dichlorethylsulphide

A similar test on a larger scale was carried out at the Edgewood Arsenal, Edgewood, Md. This was done to determine: (1) In a large group of men, the relative number of hypersensitive, average, and resistant individuals. (2) If among factory workers in mustard, the hypersensitive individuals were those most frequently burned, and the resistant those seldom burned.

The method of conducting the test was to treat each man on the forearm with a 1 percent and 0.01 percent solution of mustard gas in paraffin oil. Twenty-four hours later the man was examined for a positive or negative reaction to these solutions (indicated by the presence or absence of visible erythema). In a group of 915 men, the actual time for 6 workers to apply the solutions was about 10 minutes, and to read the tests the next day about 20 to 25 minutes, including taking the names of the hypersensitive men.

Table 32 indicates the relative number of resistant, average, and hypersensitive individuals.

TABLE 32.- Effect in individuals of local application of dichlorethylsulphide

The number of men available for testing the second proposition was too small to arrive at definite conclusions (only 31 men from the mustard-gas plant were present). Tests were also made on 51 hospital cases suffering from eye or skin effects of mustard gas. It was evident from these that average and resistant men were frequently burned. An examination into conditions at the plant showed that most of the casualties had occurred in cleaning pipes and where breakage occurred, under conditions where any man would be burned. Any individual will be burned from the application of the liquid to the skin or from a high enough vapor concentration. It was true, however, that where a sensitive and a resistant individual were burned under the same conditions the sensitive one had the more severe burn and required longer treatment.


In cases of relatively long exposure to Vapor of fairly low concentration, it is certain that conditions will arise where hypersensitive men will be burned severely enough to become casualties, while resistant and average individuals will escape without burns or only slight ones.

Table 33 is the result of a test made on negroes. The few of these previously examined by the vapor test had been found very resistant. Only the 1 percent and 0.1 percent solutions were used.

TABLE 33.- Sensitivity of negroes to dichlorethylsulphide

It is seen from Table 33 that negroes as a race have a much more resistant skin than white men. No negro of the 84 examined reacted to the 0.1 percent solution, and of course none would react to a more dilute one. About 10 percent of white men reacted to the 0.1 percent solution, while 2 to 3 per cent reacted to the 0.01 percent solution, or were hypersensitive. About 78 percent of the negroes failed to react to the 1 per cent solution, while only 20 to 40 percent of the white men did not show a reaction.


In the above discussion it has been tacitly assumed that the sensitiveness of a single individual is always the same. Such is not the case. The sensitiveness of a single individual is the same from day to day provided all conditions are the same, but may vary greatly under different conditions-the physiological variation in sensitiveness of the individual.

In order to show that the sensitiveness of the individual does notary from day to day under the same conditions, a number of determinations of susceptibility were made upon the same individual on successive weeks. These determinations showed that the sensitiveness of an individual did not vary greatly, probably less than 25 percent from day to day. An individual who was positive to a 15-second exposure on one day could be negative to 15 seconds but positive to 30 seconds a week later. He was never negative to a 1 or 2-minute exposure. Also, an individual who was negative to a 4-minute burn and positive to a 5-minute burn at one time could be negative to a 3-minute burn and positive to a 4-minute burn a week later, but he would not react to a 1 or 2-minute burn. So constant, in fact, was the sensitiveness of an individual that it was possible to plot a curve of the least time required to produce a burn (erythema) at different concentrations. If the variations in sensitiveness from day to day had been great, it would have been impossible to plot these curves, and there would have been many burns on the wrong side of the curve, and many negatives on the other side.


There are undoubtedly many conditions which influence the sensitiveness of the individual to mustard gas, but of these only two have been thoroughly investigated; sweating and moisture. A number of individuals from the laboratory were selected and given burns upon the forearm of different lengths of time, from 1 minute to 5 minutes, according to the known sensitiveness of the individual. A very mild burn was desired. The method used for burning was the "standard test-tube vapor test.'" The individual was then instructed to go out and run until the whole body, including the forearm, was in a profuse sweat. He was then given two other burns of the same length exposure, as before, one upon the moist, sweating arm, the other upon a part of the arm which was wiped dry with a towel. It was noted that in every case both of the burns made after exercising were distinctly more severe than those made before exercising, and the skin which was wiped dry usually received a more severe burn than the skin which was left moist. In another experiment performed in a similar way the feet of the subject were kept in a tub of hot water until the whole body was in a profuse sweat. The results were essentially the same. This action may have been due to the effect upon the sweat ducts, as it is known that the gas enters most readily these ducts. On the other hand, the action may have been due to the moisture. Experiments were carried out on the effect of moisture in the following way. An area of the forearm was kept moist for a few minutes with wet cotton. The sponge was then removed, and two vapor tests were made, one over the moist area and one over the normal dry skin. In the first three cases mild burns were given; in the last three cases, severe burns. The moist burn was always the more severe, in one case producing a blister when the other did not.


The paraffin oil test was used on a number of animals and indicated that differences in susceptibility existed in different species and in different individuals of the same species. Table 34 shows the results.

TABLE 34.- Susceptibility of skin of animals to dichlorethylsulphide

The horse appeared to be the most sensitive and the monkey and guinea pig the most resistant species, while the dog seemed to have a sensitivity as near man as any of the other species. However, the number of animals examined was too small for any far-reaching conclusions, and it is to l)e noted that no animal exhibited a reaction at all similar to that in man. No animal has yet been found which will give a blister from the application of mustard gas.



The only factor in the general characteristics of the skin which would appear to have a distinct bearing on the question of sensitivity is the apparent thickness of the skin. The fact that negroes as a race are much less susceptible to this substance than the white race furnishes a clue to the reason for susceptibility. Any constant differences in the skin of negroes and white men might furnish a basis for future work.

The following experiments tend to throw some light on the mechanism of absorption of mustard gas by the skin and, hence, on the question of susceptibility.

Fifteen minutes after pure dichlorethylsulphide was applied to the skin it could be almost entirely removed by long-continued rubbing with kerosene. This indicated that the substance was not immediately absorbed into the deeper layers of the skin. A resistant individual whose normal minimum burning time was 4 to 5 minutes and who did not react to the 1 percent oil solution, could be made to give a reaction to a 15 to 30-second exposure or a 0.01 percent oil solution by covering the area of application with a glass cup immediately after and leaving it covered for 6 to 8 hours.

It was found that when two vapor tests were made under identical conditions on the arm of a sensitive individual (5 minutes' exposure) and immediately after the exposure the arms of the two individuals were impressed on each of the exposed areas for 5 minutes, the burns were modified in intensity. If the recipients were respectively more or less resistant than the donor, that burn on the donor's arm given to the more resistant man was the mildest. This is very striking. If a sensitive individual impresses his arm alternately against burns of the same concentration and exposure on a resistant and sensitive man, the recipient receives a more severe burn from the sensitive than from the resistant man. This would indicate that the skin of a resistant individual displays a greater affinity or capacity for dichlorethylsulphide than that of a sensitive man. A tentative explanation of this phenomena can be made as follows: A three-phase system is involved-the air over the skin surface constitutes the outer phase; some fatty or keratinous elements of the skin the central phase; and a cellular portion of the skin the inner phase. The central phase is rich in lipoids and poor in water, while the inner phase is rich in water and poor in lipoids. After exposure to the vapors of dichlorethylsulphide the central phase is the absorbing agent and tends to establish equilibrium with the other two phases. On account of the lipoid nature of the central phase no damage is produced here because the compound is not hydrolized. On its passage from the central to the inner phase, hydrolysis takes place within the cell and damage results when a sufficient concentration of hydrochloric acid is attained. The outer phase is constantly being freed from vapor by diffusion and convection currents, so more and more can evaporate from the central phase. The susceptibility of an individual depends on the relative power of the central phase to hold the poison in an inactive form (not hydrolyzed) and prevent its entry into the inner phase at a sufficient velocity to result in the formation of a toxic concentration. No attempt was made to localize the central or inner phase with any definite structure of the skin. As mustard gas is known to penetrate the sebaceous ducts, the fat here might form one phase and the epithelial lining another.



Anyone who observes the tremendous difference between individuals is struck with the practical use that might be made of this difference. There are individuals who show distinct burns (erythema) from a 5-second exposure, while working beside them are men who require an exposure of 5 to 10 minutes to produce a burn. The hyperemia produced by the 5 seconds' exposure is not serious when limited to a small area, where it is closely surrounded by normal tissue, but a hyperemia of no greater intensity when produced over a large area of the body may make it imperative that this individual be relieved from duty. The edema and itching may be quite distressing, and prevent the patient from sleeping or resting. He may have to spend most of his time bathing the injured parts with cold water to get relief. The man beside him, however, may be exposed to the same concentration for 10 to 50 times as long and show absolutely no effect. The former individual should not be allowed to work in any region where there is even a low concentration of mustard gas, for if he does he is certain to be burned severely, and practically certain to become a casualty. The latter individual, on the other hand, would make a splendid workman in a mustard-gas factory or a splendid soldier in a territory which has been shelled with mustard gas, if provided with a mask.

The following suggestions appear warranted from this investigation in sensitivity to mustard gas:

1. Sensitivity tests should be made routine in all plants handling mustard gas, and only the more resistant individuals should be stationed there if possible. Application of a single solution is all that is necessary. The 1 percent solution will give about 40 percent of the total men examined to be used in mustard gas. If a larger percentage is desired, lower percentage (0.5 percent) solution can be substituted.
2. The feasibility of withdrawing the hypersensitive men from the line from any duty where mustard gas may be encountered should be considered. This would involve about 3 percent of the men. The actual carrying out of sensitivity tests would be very simple-the application of a 0.01 percent solution of mustard gas in paraffin oil to the forearm of each man and the observation of this spot 24 to 36 hours later. The time required would be much shorter than that necessary for venereal inspection of a company. The hypersensitive men are certain to become casualties if exposed to mustard vapor even in low concentrations, and in this way be eliminated afterwards if not before.



At an early stage in these investigations it was observed that after exposure of the skin to either the liquid or vapor the most efficient method of removal was washing with some solvent of low volatility, such as kerosene, followed by thorough rinsing with soap and water. With such a treatment it was found possible to apply the oil to the skin and by washing after 2 or 3 minutes to either prevent the subsequent delayed reaction entirely or to reduce it to a slight hyperemia. The very narrow time limits within which these preventive measures were effective indicated a remarkably rapid absorption of at least a


part of the gas. It was found, however, that if the sponging with kerosene was continued vigorously for 20 to 30 minutes, with frequent renewals of both sponge and solvent, vesication could be prevented in most cases after an exposure of 10 minutes or less and in rare instances after 15 minutes. (Table 35.)

TABLE 35.- Removal of dichlorethylsulphide from skin by washing with kerosene

The following tests were made on the shaved and dried skin of the forearm by the "nailhead method." A nailhead 3 mm. in diameter was pressed firmly against filter paper saturated with dichlorethylsulphide and then applied with firm even pressure to the skin. After a definite period of time the exposed area was washed or "treated." This period of undisturbed contact between the oil and the skin (or, in subsequent experiments, between the vapors and the skin) will hereafter be referred to as the length of exposure. Since the reaction did not appear for several hours, and since the burns did not develop sufficiently to permit of accurate comparison in less than 24 hours, the first readings were made at the expiration of that period, and final readings after 48 hours. The reactions as observed at that time are graded as follows:

+ Mild erythema.
+ + Moderate erythema and swelling.
++ + Mild vesication with slight surrounding erythema.
+ +.+ + Severe vesication with slight surrounding erythema.
++ + + + Severe vesication with marked surrounding erythema and edema.

It should be noted in the table that + + + indicates the penetration of sufficient mustard gas to cause a blistering burn while + and + + indicate that the amount of mustard gas introduced is insufficient to cause a blister, but only redness and slight swelling. Beyond the limits given above it is unsafe to make quantitative comparisons of two or more burns.

A 10-minute exposure, if washed with kerosene for 10 minutes or less, blistered; the actual blistering could be prevented in most cases by washing for 15 minutes or more. Little added benefit could be obtained by treating longer than 25 to 30 minutes. Where the time of exposure was increased to 15 minutes or more vesication could barely be prevented, even after rubbing or sponging for an hour.


That dichlorethvlsulphide was actually being removed from the skin after the first few minutes of treatment was shown by the observation that unless the sponge and kerosene used were frequently changed, and the surrounding areas thoroughly washed, the resulting burn was greatly increased beyond the size of the original exposure--following in general the areas wet by kerosene--and in some cases even the operator's fingers became contaminated with the gas
and a burn resulted.

Prolonged washing was tried with soap and water after 10 to 15 minutes in a similar manner; but though some benefit was observed, it was not nearly as marked as when the washing was done with kerosene or some solvent for the oil (vaseline, acetone, alcohol, and benzene were tried).

From these observations, it is quite evident that the dichlorethylsulphide was at first rapidly taken up by some element in-or adjacent to-the surface of the skin, and for 2 to 3 minutes it might be completely removed, and for 10 to 15 minutes partially removed, by prolonged washing with an organic solvent, and to a lesser extent with soap and water. This conclusion has been emphasized because it is an important consideration in dealing with the ultimate absorption of the substance.

In the absence of accurate information regarding the physical and chemical constitution of the outer layers of the epidermis and the exact location and condition of the mustard gas when first taken up, it appears inadvisable to speculate as to whether the gas is at first absorbed on the skin surface and then penetrates the sweat glands by capillary or absorption, or a mixture of both, or whether it passes to a certain extent directly through the cortical layers by a process of absorption or solid solution. Substances like keratin might be expected to absorb the compound and fats and lipoids to absorb or dissolve it. It is sufficient for the moment to consider the skin as a protective medium taking up the compound on its surface and offering a definite resistance to passage of the gas inward. This resistance may be gauged by determining the amount of gas which it is necessary to apply externally to insure the delivery to the inner tissues of a concentration sufficient to produce recognizable toxic effects. The nature of this resistance will become evident after the discussion of the loss of the gas from the skin by evaporation.


It has been shown that individuals vary in their sensitivity to slight quantities of dichlorethylsulphide,11 and a method for demonstrating this variation between individuals, by determining the shortest exposure to the vapors which will produce a visible reaction, has been described. Since any individual gives an increasingly severe reaction with increasing length of exposures until the full effects of the irritant are obtained (severe vesication), there should be a definite relationship between the concentration of the dichlorethylsulphide and the length of the exposure required to produce burns of equal intensity. The method used in working out this relationship was based upon that used in "gassing chambers." 12

The principle was briefly as follows: Dry air was bubbled through dichlorethylsulphide at a known and constant rate, and mixed with a definite quantity of dry air also under constant flow for the purpose of dilution. A suitable glass exposing chamber was placed in the circuit near the bubbling tube,


using as little rubber tubing as possible for connections. (Rubber tubing absorbed the gas very readily.) The concentration could be determined by analysis of the escaping gas and air mixture from the open end of the exposing chamber, but could be accurately and more readily estimated by the loss-in-weight method, figuring the loss in weight of the bubbling tube after a certain period of time in relation to the known quantities of air employed. The skin of the forearm of the subject was shaved and sponged with alcohol and the desired exposures were made. Stress is again laid on the fact that readings were made after 24 and again after 48 hours, since the burns did not develop fully until that time. A faint disk of erythema was considered positive.

While curves drawn were apparently hyperbolic in form and consequently asymptotic, there appeared to be a critical level in concentration at approximately 0.002 to 0.005 mgm. of gas per liter at which no injurious effects were noted after prolonged exposures. From this it must be concluded that the amount of gas finding its way into the interior tissues after an exposure to the concentrations in question was no more than can be metabolized by the cells.This corresponds with previous observations regarding the threshold concentration of gas that can be disposed of without injury in prolonged respiratory tests.

It is essential to mention here that in determining the toxicity of dichlorethylsulphide on animals by inhalation a similar phenomenon was observed; data expressing the relation of time to concentration indicated a marked critical condition at this same mean concentration, 0.005 mgm. per liter. While the data for different species indicated differences in sensitivity, all indicated an ultimate threshold concentration within the limits of 0.01 and 0.001 mgm.per liter. It seems evident that this concentration represented a border line where the organism generally can metabolize the poison without serious effects.

It must be remembered that these experiments were originally carried out with a purely practical military objective: The determination of the relation between the concentration of gas and the time of exposure required to produce a burn under field conditions. It was fully realized that the loss from the skin by evaporation was considerable, and that exact scientific data regarding absorption of the gas by the skin and its penetration into the interior tissues could be secured only by covering the exposed area with some impervious substance immediately after the completion of treatment. Unfortunately, the suspension of experimental work on the cessation of hostilities prevented the carrying out of an elaborate series of experiments on these lines with varying concentrations of the compound. But from experiments with one concentration, to be reported later in this chapter, it has been demonstrated that the loss from the skin was extremely great, far exceeding the amount that penetrated the tissues, and for this reason it appears inadvisable to draw any far-reaching conclusions from the preceding curves.

The fact that a definite proportionality appears to exist between the increase in time at a given concentration and the increase in concentration at a given time required to produce equivalent effects in several individuals, suggests that the problem is one of satisfying the capacity of the skin to such an extent that the requisite amount of gas may penetrate to the interior tissues. The question as to whether this difference is due to the resistant man was made the subject of further experiments.



In considering any factor which is known to modify the final intensity of a mustard-gas burn, one must be careful to differentiate between those which modify absorption only and those which may have some effect upon the ultimate reaction which follows the introduction of the substance into those organized elements of the skin in which pathological effects are produced. Where the former may be brought about through simple physical means, the latter no doubt involve physiological relations of much more complex nature and are not dealt with in this chapter.

As previously explained, no attempt is made to specify what agencies are involved in the preliminary or final absorption of dichlorethylsulphide by the skin. The detailed localization of these agencies have not been found necessary to a clear conception of the general mechanism. Passing from the external absorbing agent previously indicated, the dichlorethylsulphide, in all probability, would pass through the deeper lying elements to the lipoid constituents of the cells of the adjacent tissues. If one were dealing with true absorption into the intermediary binder, the passage to the deeper tissues would probably be by diffusion from solid solution; on the other hand, its retention in this intermediary binder by absorption would make it highly probable that capillary and surface tension effects would direct it into the ducts, from which it would be taken tip by absorption. On reaching the lipoids of the cell wall or cell contents proper, the mustard gas would slowly diffuse into the aqueous constituents of the cell, where hydrolysis of the compound would result and its toxic action would begin. Again attention must be called to the danger of attempting to localize the site or mechanism of this action from the evidence on hand. The nature of the poison may be such as to effect a fatal anesthesia of the vasomotor nerve supply, resulting in prolonged vasodilation and subsequent edema and vessication. On the other hand, there is abundant evidence from the work done by Lillie, Clowes, and Chambers on the penetration of marine eggs to indicate that the intracellular liberation of acid which would follow the hydrolysis of the dichlorethylsulphide would cause such profound changes in the protoplasmic equilibria of the cells concerned that their normal metabolism would be permanently upset and their permeability increased.5 From the evidence on hand, one is forced to believe that equal concentrations of the poison in this last phase (inside the cells of the tissues affected) would produce similar and nearly equal reactions in almost all men. Whereas the extreme difference in sensitivities of the skin represents an order of 1 to 600, the extreme difference in toxicity for any one species (dog), by inhalation or injection, is of a much smaller order, possibly 1 to 5. Exceptions to this equality of response to equal intracellular concentrations have been noted. One case has been reported where the subject (H. W. S.) developed an abnormal reaction suggesting an anaphylactic condition.

It has been suggested that after exposure of the skin to small amounts of mustard gas, evaporation from the surface has an important bearing upon subsequent absorption. To test this point, exposures to the vapors were made with the test-tube method previously described. Two exposures, just sufficient to cause a mild reaction, were made at the same time and under identical conditions. One was immediately covered with a shallow glass thimble of approximately the same diameter as the exposed area, fastened to the arm for three


hours or more with adhesive tape. The other was left uncovered, and served as a control. After 24 to 36 hours, when both burns had reached a maximal comparative development, the covered burn was seen to be much the worse. A series of experiments performed in this manner is tabulated in Table 36.

TABLE 36.- Effects of evaporation after application of dichlorethylesulphide

Column I.-Subjects average minimum burning time to the standard testtube, 20 0 C.
Column II.-Length of exposures made.
Columns III and IV.-Relative intensity of resulting burns: + + + severe erythema; + +mild; + faint; -negative. (Blistering burns were not made in any case.)

Sensitive men (i. e., men who gave a positive reaction to short exposures) generally showed less increase in reaction than did men who were relatively much more resistant. This suggested that the minimal burning time be determined when the exposed skin was subsequently covered for a prolonged period and compared with the minimal burning time when exposures were left uncovered.

From Table 37 it is apparent that the possible reduction in minimal burning time in a sensitive individual by covering the exposure was generally less than in a resistant individual.

TABLE 37.- Effects of evaporation after application of dichlorethylsulphide

This suggests that the skin of a resistant individual absorbed more of the dichlorethylsulphide than the skin of a sensitive one, because the only other apparent wav in which covering could influence the final reaction would be by increasing the rate of passage of the sulphide from the outer layers of the skin to the inner lavers through an increase in temperature and thus raise what would


naturally (uncovered) be a subreactive concentration to a reactive one. On such a basis it would seem obvious that a sensitive man would show relatively as great an increase in intensity as a resistant man. Against the argument that a sensitive man exhibits a close approach to the maximum reaction, and therefore can not be made to show a large increment over his normal reaction (uncovered), is the fact that individuals were known who blistered to a 5-second vapor exposure, and gave severe edema to less than 1 second, and in whose history there was nothing to indicate any anaphylactic or other abnormal reaction, beyond their great sensitivity. This problem will be considered later.

Another point of interest was the time an exposure must subsequently be covered to secure the maximum reaction on any one individual, thus finding the time during which the loss by evaporation would still produce an appreciable decrease in the final intensity of the burn. An experiment to determine this is given in Table 38. Exposures of border-line intensity were made and covered immediately, and were left covered for varying intervals.

TABLE 38.- Effect of covering dichlorethylsulphide burns

In another experiment (Table 39) the approximate rates of evaporation and absorption were determined by covering mild exposures after varying intervals of time had elapsed during which the skin had been left open to the air.

TABLE 39.- Effect of covering dichlorethylsulphide burns

From these experiments it is evident that equilibrium in the skin was reached in about 45 minutes and that capping had less effect on a sensitive than on a resistant skin. This rate of attaining equilibrium must have had a direct bearing upon further absorption and, ultimately, upon sensitivity. Since the dichlorethylsulphide was at first absorbed by the surface of the skin or some superficial elements adjacent thereto, and could be lost from this medium by evaporation, the relation rate at which it passed from this intermediary binder to some other phase, in which it was firmly fixed, as against the rate at which it was lost by evaporation, would be the determining factor in the question of whether or not a slight exposure to the vapors would prove positive or negative.


Since the intensity of the burn produced by a given concentration bore a direct relation to the length of the exposure, and since in at least one case (H. W. S.) a reaction was produced within 10 minutes of exposure, it is probable that the velocity with which the substance passes from the surface to the deeper layers was such that the absorption of additional amounts was not interfered with. While the determination of the rate at which it passed from the surface inward was a problem of considerable complexity, the rate at which it was lost from the skin by evaporation was readily determined by a few simple experiments. By making a series of exposures of varying periods of time and covering certain of these exposures to prevent evaporation, it was possible to determine the time for which a short exposure must be covered to give a burn of equal intensity to an exposure of greater length which was left open to the air.

Two such experiments were made on the skin of the shoulders of two subjects (H. E. I and W. B. Mc.); the burns were covered immediately and left covered for the time given in the left-hand columns, Tables 40 and 41. An accurate representation of their relative severity after they had reached their maximum comparative development, as determined by the readings of three observers, is given. The system of nomenclature is simply an effort to give the relative severity of the burns in each set, and comparisons can not be made from one set to the other.

TABLE 40.- Effect of covering exposures of varying lengths

In the exposures capped for 2 and 4 minutes it is probable that the thimbles were not tight. The skin used was on a curved surface of the shoulder and the resulting burns are undoubtedly low in intensity through this experimental error.

TABLE 41.- Effect of covering exposures of varying lengths

Where Table 40 is a comparison of more intense burns. Table 41 is a comparison of threshold burns, i. e., where the exposure was just sufficient to produce a faint reaction.


Since a 15-second exposure could be increased by covering to the intensity of a 4-minute exposure, the amount lost by evaporation into the air must have been far in excess of that which passed into the deeper tissues. Experimental evidence was insufficient to determine the relative distribution between air and the skin. The practical importance of the fact of evaporation can not be overlooked, and its bearing on variations in cutaneous sensitivity will now be considered from another point of view.


An interesting phenomenon was observed when the untreated normal skin of one subject was impressed for 5 minutes upon an area of skin of another subject previously exposed to the vapors of dichlorethylsulphide. Under these circumstances, both donor and recipient might develop burns (due to the transportation of the poison from one skin to the other) the intensity of which would vary according to the circumstances and the respective sensitivities of the participants. The degree of transportation was most strikingly observed in the intensity of the burn on the donor's arm. If two similar exposures were made on the arm of a sensitive man, and one of these burns was treated, so to speak, by contact for 5 minutes with the skin of a resistant man, the treated burn would be markedly less severe than the control, in some cases being entirely prevented. If, however, the recipient was equally sensitive to or more sensitive than the donor, the burns on the latter exhibited far less difference. Both treatments could be effected at once, using two recipients, one more resistant, and one less resistant, than the donor. In such a case the burn brought in contact with the more resistant skin was the less severe.

A few experiments of this nature are given in Table 42. The subjects are divided for clarity into two classes, (R) resistant and (S) sensitive. It is evident that the "resistant skin" had reduced the burn with which it was in contact much more than had the "sensitive skin." One is forced to the conclusion that it had actually absorbed more of the gas.

TABLE 42.-Transfer from skin to skin of dichlorethylsulphide burns

This same experiment was carried out in the reverse order. Similar exposures were made on the arms of a sensitive and a resistant subject. Immediately after the exposure, a third subject received both burns by impressing them against his arm for 5 minutes. It was seen that a "sensitive skin" would take more of the poison from a "skin of equal sensitivity" than from a more resistant one.


Both experiments were, in fact, important evidence that the skin of a resistant individual exhibited a greater affinity or capacity for dichlorethylsulphide than that of a sensitive one. Whatever the arrangement of the experiment, the results indicate that there was an actual partition of the gas between the two skins, with an evident tendency to establish an equilibrium in which the largest portion of the gas would remain in that skin which possesses the greater capacity for it. The experiment confirms in a striking manner the observations noted above, and explains why equilibrium with the air is attained from a sensitive skin sooner than from a resistant one. This is directly contrary to the general belief that a resistant individual is more resistant to low concentrations of mustard gas than his fellows because his skin absorbs less. The assumption that his skin absorbs more makes it imperative that some explanation accounting for the paradox be made.

In order to determine the persistence of the gas on the surface of the skin after exposure, an experiment was made in which a series of 5-minute exposures were capped for 15 minutes, 30 minutes, 45 minutes, etc., and then impressed on the arm of a subject of about the same sensitivity for 5 minutes.

Since the recipient failed to develop a burn after contact with the burn on the donor, which had been capped for 60 minutes, it was apparent the gas had disappeared from the donor's arm in that period. This agreed well with the results obtained by capping in Table 39, in which a maximum absorption was indicated within that time.

TABLE 43.- Sensitivity tests of dichlorethylsulphide burns

TABLE 44.- Persistence of gas after dichlorethylsulphide burns


It is generally recognized that substances possessed of the capacity of readily penetrating protoplasm are almost invariably soluble to a certain extent in water and are also " lipoid soluble," that is, soluble in fats and organic solvents such as benzene and xylene, and that the so-called partition coefficient between water and benzene or xylene is a factor of considerable importance in conditioning their power to penetrate living cells.

Some 25 compounds were investigated, so-called war gases, which were not only highly toxic, but which irritated the skin. The determination of their partition coefficients was a very difficult matter in most cases because


of their rapid hydrolysis in water (in such cases a strong acid is one of the decomposition products), but it was evident that they were all soluble in both lipoid solvents and water to some extent. This fact, with the fact of their hydrolysis, suggested that the mechanism of their action could be correlated in a general way with that of dichlorethylsulphide. Further experimental work is necessary to establish this fact. This group of compounds in the category of war gases vary greatly in their degree of activity and in their specific toxic effects, and while no far-reaching generalization on this question will be attempted it is proposed at a subsequent stage in this chapter to discuss briefly certain cases in which toxic effects correlate in a measure with physico-chemical properties.

A series of experiments was performed with a number of organic bases in an attempt to determine their value as a means of counteracting the effects of mustard gas. It was found that with the exception of ammonia only those bases which were lipoid-soluble irritated the skin, indicating penetration. On the other hand, too high a partition coefficient (too low solubility in water) appeared to diminish the irritating activity on the skin. (The following examples are representative of the groups tried: Ammonia, propylamine, ethylamine, amylamine, di-isoamylamine, bornylamine.) Whatever the interpretation of the details, it is apparent that lipoid solubility is an important factor in the penetration of the intact skin.

A brief reference should be made to a phenomenon requiring further investigation. Field observers have noted that burns occur more frequently on moist than on dry polrtions of the body. This observation was confirmed in the laboratory and, furthermore, a film of water on the dry skin was found to facilitate the passage of mustard gas, showing that the effects observed in the field were not attributable simply to a possible higher permeability of freely perspiring areas, but was in some way attributable to the presence of water on the skin.

To determine whether this was attributable to surface effects caused by the presence of a film of fluid, a series of equal threshold burns were made over the dry skin and skin wet with various organic agents, care being taken in the case of each individual to select the exact time of exposure required to give a mild burn on the dry skin.

The fact that water is a very poor solvent for mustard gas and yet appears to facilitate the passage of the substance into the tissues as well as the other fluids which are good solvents for it, suggests the probability that capillarity rather than solution may play an important role in transporting mustard gas from the atmosphere to the point of entry into the skin, and lends some support to the view that mustard gas passes down the sweat glands by a process in which surface phenomena play an important part.

There is a striking correspondence between the above results in which a water film facilitates the passage of mustard gas into the tissues and the observation of Clowes, Perrott, and Gordon that the passage of mustard gas through clothing is facilitated by the presence of from 3 to 5 per cent of water.13


A variation in sensitivity of the skin of several hundred to one was observed in experiments with saturated vapor or paraffin oil solutions of dichlorethyl-


sulphide. In experiments with different concentrations and times of exposure, using different species of animals, the variation in the susceptibility of the individuals of a given species to the effects of inhalation were of a very much lower order of magnitude-probably not more than five to one. Furthermore, on intramuscular or intravenous injection in dogs variations of no great order of magnitude were observed. As pointed out in a previous portion of this chapter, it is evident that the differences in sensitivity observed were due to differences in the relative amounts observed compared with the amounts lost by evaporation. Hence one is justified in assuming that in individuals of the same species the amount of mustard gas required within the cell to produce pathological effects is roughly the same-that is, the threshold concentration for the cell varies very little in different men.

In discussing the mechanism of absorption of mustard gas by the skin it is proposed to consider the problem from the standpoint of a three-phase system in which the outer phase (A) represents the external atmosphere containing varying concentrations of mustard vapors; the middle phase (B) represents the outer layers of skin through which the mustard must pass in order to produce a toxic effect; and the inner phase (C) represents the inner layer of the skin, particularly the protoplasmic contents of those cells of the tissues which are directly affected by the poison.

For purposes of convenience A, B, and C in resistant individuals have been designated as Ar, Br, and Cr, and in sensitive individuals as A8, B8, and C8.

If, as has been indicated above, the velocity with which mustard gas must pass into those reactive cells which have been designated as C in order to produce a given pathological effect, is approximately constant in resistant and sensitive individuals-if the critical toxic threshold concentration in Cr is the same as in C8-it necessarily follows that variations in concentration of gas or time of exposure in A required to produce equal effects in Cr and C. must be attributable in great part to variation in the resistance offered by Br and B8 to the passage of the gas.

The explanation for the difference in facility of passage through Br and B8 may lie in one or more of the following causes: (1) Variation in facility of absorption of mustard gas on interface between A and B. (2) Variation in the facility with which absorbed mustard gas passes along gland surfaces by capillarity or into and through the epidermis by a process of diffusion of the absorbed or dissolved gas. (3) Variation in the thickness of Br and B8 or the actual distance to be traversed from A to C. (4) Variation in the amount of gas required to saturate constituents of the skin which may be capable of adsorbing or dissolving large quantities of the gas (for example, lipoids, pigment granules, etc., in negro and white skin). (5) Variation in the facility of passage of gas from B into C particularly from the lipoid nonaqueous phase into the aqueous phase of the protoplasm of those reactive cells in which the pathological effects are produced.

It is difficult to determine the relative importance of the r6le played by each of these factors, but variation in the resistance of B may be conveniently represented pictorially either by varying its thickness, making Br thicker than B8, or by varying the angle of the average gradient of mustard-gas concentration through B required to deliver a given amount of mustard gas in C; in which case that of Br would be steeper than B8.


It will be seen that in either case, a critical concentration of mustard gas would be delivered into C. with greater facility than into Cr; thus a shorter exposure or a lower initial concentration would be required to effect the passage into C, of a reactive quantity of the poison.

Whatever mode of expression be adopted, the effect of evaporation after the exposure is evident. The concentration at the interface AB would be higher than at the interface BC, but it would rapidly decrease as a result of loss into the air.

The delivery of the necessary critical concentration of gas into C for a sufficient period of time to cause pathological changes would obviously depend upon the thickness of B, upon the angle of gradient of resistance to passage through B, and upon the facility with which the concentration falls off at the interface AB as a result of evaporation. It appears desirable to utilize the available experimental data in an attempt to determine, more specifically the relative importance of the individual factors enumerated above, viz, surface adsorption, resistance to passage of the gas, thickness of the skin, saturation capacity and threshold concentration required to produce toxic effects in C.

This last factor- i. e., threshold concentration in C-must be very low because (1) concentrations in the range of 0.002 to 0.005 mg. per liter represent not only critical points in the curves determined on the skin but (2) also correspond very closely with the threshold concentrations observed in respiratory experiments, and (3) the enormous proportion of gas lost from the skin after exposure, as shown by capping experiments, indicates that the amount which is absorbed is very small. The extremely small amounts to which some individuals give a reaction is further evidence for this fact. This low threshold concentration in C leaves adsorption on the skin surface or one or more of the factors involved in the passage through B, as primarily responsible for the buffer effect exerted by the skin.

In the case of liquid burns or severe vapor burns the difference in the protective mechanism of resistant and sensitive skins appears to be reduced to a minimum; all the factors involved are probably saturated, whatever their capacity.

The observation that the difference between Br and B8 is more and more accentuated with reduction in time of exposure or concentration points to surface adsorption as an important factor; reasoning in this direction is supported by the following facts:

(1) Ten minutes after exposure of the skin to liquid mustard gas so large a proportion of the mustard gas may be removed by repeated washings with kerosene as to reduce the amount passing into the tissues below the critical level required to produce a blistering burn. (2) Washing with soap and water produces a corresponding effect, but to a much less degree. (3) The presence of mustard gas on the skin surface may be demonstrated half an hour after exposure by transmitting a burn to another individual. (4) Since a 15-second exposure may be raised by capping for an hour to the magnitude of a 3-minute exposure, the loss from the skin to the atmosphere must be far in excess of the amount passing into the interior. This suggests that the maximum concentration of mustard is on or near the surface. (5) The curves covering these capping reactions support the theory that surface adsorption is the factor of primary importance. (6) The fact that the skin may be left open to the atmosphere


for a prolonged period, in some cases as much as 30 minutes after exposure, and the intensity of the burn may still be raised by capping, indicates that loss of gas from the surface is still taking place. (7) The affinity of the skin of the resistant individual for mustard gas is far in excess of that of a sensitive one. The fact that Br withdraws gas from B8 shows that Br adsorbs gas more strongly than B. and that the gas is still on or near the surface. (8) If exposures on resistant and sensitive skins have been so adjusted as to give equal burns if subsequently left uncovered, capping the exposures gives a worse burn in the resistant than in the sensitive case. This result corresponds with the fact that gas may be withdrawn by a resistant skin from a resistant skin longer than from a sensitive skin.

From the above considerations it seems probable that difference in the adsorptive capacity of the skin surface is the most important limiting factor in determining the degree of exposure in A necessary to effect the delivery of a toxic concentration in C.

In considering the relative importance of factors other than the surface adsorption on B and threshold concentration in C, it is obvious that the greater resistance of the negro's skin as compared with the white's skin may reasonably be attributed, in part at least, to differences in actual thickness of the skin, to fats and lipoids which appear to be more plentiful and to pigments which are known to be more plentiful in the negro skin than in white.

The very short minimum time within which the first reactions become visible in certain sensitive individuals would seem to indicate that the first traces passed through the skin fairly rapidly.

The experiments with water films seem to support the view that mustard gas passes into the sweat glands from which it is adsorbed after passing by capillarity to areas adjacent to the cells of the underlying tissues in which pathological effects are produced.

The capping experiments support this view. The passage of mustard gas into the sweat glands would obviously be facilitated by maintaining the vapor pressure.

These experiments should be considered in the light of certain observations concerning the absorption of dichlorethylsulphide in other parts of the body.

Even though mustard gas is very rapidly hydrolyzed at body temperature, inhalation of high concentrations of the vapor, intravenous injection of large quantities of the saturated aqueous solution, and application of the oil to the skin, cause not onlv local lesions but also marked and characteristic systemic effects.

It is obvious, therefore, that after its introduction into the blood stream a portion at least of the unhydrolyzed material must be taken up by lipoidal constituents and thus protected from immediate hydrolysis. After intradermal, subcutaneous, intraperitoneal or intramuscular injections of this substance the severe necrotic lesions which might be expected from its action are not observed. The pathological changes consist largely in the development of intense edema at the site of injection, with the characteristic effects on the gut and adrenals. It is possible that, when it is thus injected directly into the deeper tissues which are well supplied with blood, and which lack the nonaqueous binder of the skin, it is rapidly absorbed without doing much local damage. Moreover, the bountiful blood supply would go far to counteract the intracellular liberation of acid, and would thus circumvent its local toxic action. This suggestion is supported by the pathological changes ol)served in the respiratory tract in animals


gassed with mustard gas. Though congestion and necrotic sloughing appear in the upper tract, the lower tract is practically undamaged until very high concentrations are used, when some edema may appear. Since the lower respiratory tract is primarily an absorbing tissue, it is in very close association with the blood stream. The absorbed poison might be rapidly removed, and such hydrolysis as would inevitably occur might be counteracted by neutralization in the buffer system of the blood.

In considering the toxic action of a series of war gases it is observed that there is marked contrast in the pathological effects exerted upon the upper and lower respiratory tract and upon the skin.

There appears to be a certain relation between the areas inhibiting maximum pathological changes and certain physical and chemical properties of the gas, for example, vapor pressure, lipoid-water distribution coefficient and rate of hydrolysis in water.

Phosgene and superpalite, which have a very high vapor pressure and which hydrolyze very rapidly on contact with water, exert a very destructive effect on the lower respiratory tract, but do not markedly affect the upper tract or irritate the skin.

It is obvious that under a given condition of vapor pressure, lipoid solubility and other factors, a given amount of gas finds its way into the cell in a given time and if the rate of transportation through the cell is very rapid, the question as to whether or not a toxic concentration of acid will be produced within the cell depends upon the rate of hydrolysis, or the relation between transportation velocity and rate of hydrolysis.

A low lipoid-water partition coefficient should facilitate the decomposition of gas within the cell by raising the concentration in the water phase and by diminishing the proportion removed by the lipoid constituents of the blood.

The failure of such volatile substances as phosgene and superpalite to exert any effect on the skin is readily explained in the light of experiments with mustard gas, the loss of which from the skin to the air has been demonstrated to be very great in spite of its relatively low vapor pressure.

Substances like mustard gas, with a low vapor pressure and which hydrolyze comparatively slowly, would, after their absorption by the lower respiratory tract, have an opportunity to pass into the lipoidal phases of the blood before hydrolysis reached such dimensions within the cell as to cause pathological effects.

But in the cells of the upper bronchi and the skin, where transportation is comparatively insignificant, an accumulation would result, and ultimately sufficient acid would he liberated to produce toxic effects.

Phenyldichlorarsine and ethyldichlorarsine, with a low vapor pressure (resembling mustard gas somewhat in their physical properties) but having a high rate of hydrolysis, attack the skin and upper respiratory tract just as mustard gas does, and cause profuse edema in the lower respiratory tract just as do phosgene and superpalite.

Unfortunately, these investigations were incomplete when the suspension of hostilities brought an end to experimental work.

Consequently, though a large amount of additional data is available suggesting a relation between the physicochemical and pharmacological properties of substances which played a rôle in chemical warfare, it does not appear advisable to attempt further generalization at the present stage of our knowledge.



The experimental data indicate -
1. That mustard gas is first absorbed by some element on or immediately adjacent to the skin surface.
2. That while a portion of the mustard gas passes rapidly inward to a point from which it can not subsequently be removed the greater portion remains on or near the surface for a considerable period, a proof of which is that it may be removed even after 10 or 15 minutes by persistent washing with organic solvents.
3. That the amount of mustard gas passing into the atmosphere from an exposed surface far exceeds the amount passing into the inner strata of the skin. This loss is very great at first and is still demonstrable after 45 minutes.
4. That the time of exposure necessary to produce a positive reaction bears a definite relation to concentration and varies for different individuals.
5. That a resistant skin absorbs far more gas than a sensitive skin, and gas may be withdrawn from the latter by the former. That difference in sensitivity of different skins is due principally to difference in saturation absorptive capacity. 6. That the intracellular threshold concentration of gas required to produce pathological changes in the skin is approximately the same in resistant and sensitive individuals.


(1) Harvey, E. N.: The Permeability of Cells for Acids. Publication No. 212. Carnegie Institute of Washington, D. C., 1915, viii, Part vi, 132.
(2) Sanno, Y.: Ueber den Einfluss der Temperatur auf die Giftempfindlichkeit des Frosches. Archiv für experimentelle Pathologie und Pharmakologie. Leipzig, Aug. 7, 1911, lxv, 325.
(3) Meyer, Victor: Physiologische Wirkung der gechlorten Schwefelathyle. Berichte der deutschen chemischen Gesellschaft, Berlin, 1887, xx, No. 2, 1729.
(4) Warthin, A. S., and Weller, C. V.: The Pathology of the Skin Lesions Produced by Mustard Gas. The Journal of Laboratory and Clinical Medicine, St. Louis, Mo., 1918, iii, No. 8, 447.
(5) Lillie, R. S., Clowes, G. H. A., and Chambers, R.: On the Penetration of Dichlorethylsulphide (Mustard Gas) into Marine Organisms, and the Mechanism of its Destructive Action on Protoplasm. Journal Pharmacology and Experimental Therapeutics, Baltimore, Md., 1919, xiv, No. 2, 75.
(6) Graham, E. A.: Late Poisoning with Chloroform and Other Alkyl Halides in Relationship to the Halogen Acids Formed by Their Chemical Dissociation. Journal Experimnental Medicine, New York, 1915, xxii, No. 1, 48.
(7) Nef, J. U.: Dissociationsvorgange bei den Alkylathern der Salpetersaure der Schwefelsaure und der Halogenwasserstoffsauren. Annalen der Chemie, Berlin, 1899, cccix, 126.
(8) Weber, S.: Ueber die Giftigkeit des Schwefelsauredimethylesters (Dimethylsulphates) und einiger Verwandter Ester der Fetreihe. Archiv für Experimentelle Pathologie und Pharmakologie, Leipzig, December 19, 1901, xlvii, 113. Also:  Michiels, J.: Sur la Toxicite du Sulfate Neutre de Methyle. Archives Internationales de Pharmacodynamie, Paris, 1911, xii, 467.
(9) Erdmann, P.: Ueher Augenveranderungen durch Dimethylsulfat. Archiv für Augenheilkunde, Wiesbaden, 1908-09, lxii, 178.
(10) Weber, S.: Loc. cit.
(11) Meyer, Victor: Ueher Thiodiglycol Verbindungen. Berichte der deutschen chemischen Gesellschaft, 1886, Berlin, xix, No. 3, 3259.
(12) Marshall, E. K., jr., and Kolls, A. C.: An Apparatus for the Administration of Gases and Vapors to Animals. Journal of Pharmacology and Experimental Therapeutics, Baltimore, Md., 1919, xii, No. 8, 385.
(13) Personal communication from Dr. G. H. A. Clowes in possessions of the author of this chapter.