UNIVERSITY OF CALIFORNIA PUBLICATIONS COLLEGE OF AGRICULTURE AGRICULTURAL EXPERIMENT STATION BERKELEY, CALIFORNIA CONTROL OF GROUND SQUIRRELS BY THE FUMIGATION METHOD BY G. R. STEWART AND JOHN S. BURD BULLETIN No. 302 December, 1918 UNIVERSITY OF CALIFORNIA PRESS BERKELEY 1918 Benjamin Ide Wheeler, President of the University. EXPERIMENT STATION STAFF HEADS OF DIVISIONS Thomas Forsyth Hunt, Director. Edward J. Wickson, Horticulture (Emeritus). Herbert J. Webber, Director Citrus Experiment Station; Plant Breeding. Hubert E. Van Norman, Vice-Director; Dairy Management. William A. Setchell, Botany. Myer E. Jaffa, Nutrition. Charles W. Woodworth, Entomology. Ralph E. Smith, Plant Pathology. J. Eliot Coit, Citriculture. John W. Gilmore, Agronomy. Charles F. Shaw, Soil Technology. John W. Gregg, Landscape Gardening and Floriculture. Frederic T. Bioletti, Viticulture and Enology. Warren T. Clarke, Agricultural Extension. John S. Burd, Agricultural Chemistry. Charles B. Lipman, Soil Chemistry and Bacteriology. t Clarence M. Haring, Veterinary Science and Bacteriology. Ernest B. Babcock, Genetics. Gordon H. True, Animal Husbandry. James T. Barrett, Plant Pathology. Fritz W. Woll, Animal Nutrition. Walter Mulford, Forestry. W. P. Kelley, Agricultural Chemistry. H. J. Quayle, Entomology. J. B. Davidson, Agricultural Engineering. Elwood Mead, Rural Institutions. H. S. Reed, Plant Physiology. James C. Whitten, Pomology. fFRANK Adams, Irrigation Investigations. C. L. Roadhouse, Dairy Industry. Frederick L. Griffin, Agricultural Education. John E. Dougherty, Poultry Husbandry. S. S. Rogers, Olericulture. R. S. Vaile, Orchard Management. J. G. Moodey, Assistant to the Director. Mrs. D. L. Bunnell, Librarian. DIVISION OF AGRICULTURAL CHEMISTRY John S. Burd. Paul L. Hibbard. {Guy R. Stewart. Walter H. Dore. Dennis R. Hoagland. t Arthur W. Christie. James C. Martin. t In military service. f In co-operation with office of Public Roads and Rural Engineering, U. S. Department of Agriculture. CONTROL OF GROUND SQUIRRELS BY THE FUMIGATION METHOD By G. B. STEWART and JOHN S. BURD In the present war emergency the stimulation of food production has become the active aim of everyone connected with agriculture. It is evident that anything which prevents loss of food after it is raised is just as important as an increase in the annual crop. Prob- ably the greatest preventable loss of foodstuffs in the State of Cali- fornia is due to the depredations of the common ground squirrel. This animal is estimated by various authorities 1 to consume from five to twenty million dollars worth of food and forage annually in California. In some of the larger single counties the loss is said to reach five hundred thousand dollars yearly. This explains the present active interest in ground-squirrel control. The present investigation was undertaken to aid this emergency work and to assist in the cam- paigns carried on by the Horticultural Commissioners and by the Farm Bureaus of the state. The five most important methods of ground-squirrel control, as pointed out by Dixon in Circular 181 of this station, are : " (1) poison- ing with strychnine; (2) fumigation with carbon bisulfide; (3) trap- ping; (4) shooting; (5) encouragement of the natural enemies of the ground squirrel." The first of these methods, poisoning with strych- nine-coated barley, is undoubtedly the most satisfactory and the least expensive to use for a preliminary treatment. The work of the United States Biological Survey, the United States Public Health Service, and of several state experiment stations, has shown that the intelligent distribution of poisoned barley will remove approximately ninety per cent of the squirrel population. Reducing the remaining ten per cent should be the constant aim of effective squirrel control. The females of this small fraction will produce an average of eight young in a litter. If these young are added to the mature survivors there will be at least half as many squirrels to eat the crop as were present before the poisoning started. It can be readily seen that a follow-up treatment is essential after each poisoning campaign. The aim of squirrel extermination should be to remove all possible survivors before the breeding season. One female killed in March may remove as many as nine from the grain fields the next July. 208 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION The best method which has so far been devised for removing the remaining squirrels is fumigation with carbon bisulfide. This does not depend on the co-operation of the squirrels as does poisoning. When green food is abundant it has been noticed that many squirrels will not eat the poisoned grain. Others are especially strong and vigorous and so escape the action of the poison. With carbon bisul- fide all that is necessary is to bring the squirrel into an atmosphere which contains approximately 2% of the gas and keep it there for at least twenty minutes. Carbon bisulfide is best used when the ground is moist. The soil does not then absorb the gas in the open pore spaces and the cracks are closed through which it might escape. If carbon bisulfide is used at other seasons of the year either larger quantities must be employed or special apparatus must be used which will pump the gas through the burrow with sufficient speed to send it to the nest chamber before it has a chance to be absorbed by the soil. CARBON BISULFIDE AS A FUMIGANT The first person to use carbon bisulfide as a method of squirrel control was Dr. E. W. Hilgard, for many years director of this station. In 1876 2 he advised its use upon squirrels because he had noted it was successfully used in France to exterminate rats in sewers. Two years later 3 he described his work on the university campus, and the success- ful manner in which he had removed the greater part of the squirrels from the gardens. Since that time three general methods of use have been developed. The first is that originally recommended by Dr. Hilgard. One to two ounces of carbon bisulfide are poured on cotton waste and placed deep down in one or two of the most used entrances of each burrow ; all holes are closed with clods of earth or sod. These are inspected the following day to observe whether the treatment has been effective. The few which have been opened by the squirrel are re-treated with a slightly larger dose and inspected later. This is commonly termed "waste ball" method of treatment. Another very effective procedure has been developed by the United States Public Health Service. In it a machine called a "destructor" 4 is employed to drive the carbon bisulfide gas into the further ends of the burrow. The detailed plan of this instrument was reproduced in Circular 181 of this station. Briefly, it consists of a pump of large bore which forces a strong current of air through a chamber in which a measured quantity of carbon bisulfide is vaporized. The outlet hose of the machine is thrust down one of the principal entrances of the CONTROL OF GROUND SQUIRRELS BY THE FUMIGATION METHOD 209 squirrel burrow, the soil is filled in around it and all other entrances are closed. About fifteen double strokes of the pump are sufficient to fill the burrow with a deadly quantity of gas. One or two cheaper machines have been devised, but so far as the writers have been able to learn, neither these nor the "destructor" are at present on the market. This is due partly to the present high cost of the materials employed and partly to a lack of appreciation of the greater efficiency which is achieved by the use of a pump machine. The other variation in the use of carbon bisulfide consists in explod- ing the gas in the burrow. Carbon bisulfide is extremely inflammable and very explosive. The common procedure is to place the waste ball saturated with carbon bisulfide in the burrow, close the entrance with a loose clod of earth for two or three minutes to allow the gas to vaporize, then to remove the clod and stepping back and to one side to apply a small lighted torch consisting of a cloth soaked in kerosene wrapped around a thin piece of pipe. The explosion will be extremely vigorous and has the advantage of indicating whether all the entrances of the burrow have been previously closed. This method of operation should never be employed when the grass is at all dry. Many serious fires have been caused in this way during the drier periods of the year. All three methods of fumigation have had their advocates in the past. There is good evidence from the figures of Surgeon Long 4 that the treatment by machines of the destructor type is the most eco- nomical. He found that 200 to 250 holes per gallon could be treated with the destructor, against 50 to 60 holes by the waste ball method. The question of the relative effectiveness of the exploded as com- pared to the unexploded gas depends upon two factors, viz., whether the exploded or the unexploded gas is the more poisonous, and the completeness with which either gas is distributed into the more dis- tant parts of the burrow. If the unexploded and exploded gas are equally poisonous the more effective of the two methods is the one which will distribute the gas the more quickly and thoroughly. If the unexploded and exploded gas are not equally poisonous the more poisonous form may be the more effective even if its dissemination is slower. A -knowledge of the kinds and amounts of gas resulting from the explosion of carbon bisulfide and a comparative study of their toxicity to squirrels is therefore an important part of the present investigation. It has been generally assumed that the products of the explosion of carbon bisulfide in an excess of air consist exclusively of carbon dioxide and sulfur dioxide. This assumption is not justified, because the course of the reaction may be modified by several factors. In 210 UNIVERSITY OF CALIFORNIA — EXPERIMENT STATION the field the conditions which are most likely to affect the explosion and change the products of combusion are, the humidity of the air, the temperature, and the atmospheric pressure. It is also possible that the products formed from gaseous carbon bisulfide will be different from those produced from burning the liquid material. We have found in laboratory experiments reported elsewhere 5 that where carbon bisulfide is completely vaporized and exploded in the presence of an excess of air, the largest portion of the gas (40% to 60%) went to form carbon dioxide and sulfur dioxide, from 25 to 35% formed carbon monoxide and sulfur dioxide, and 15 to 30% remained unexploded. The carbon dioxide formed is well known to be non-poisonous. It is the gas used to carbonate soda water and similar beverages. The only way that it could cause death would be if it were present in such large amount that it excluded the air. The squirrel would then die from suffocation ; this would require a vastly greater quantity than is ever present from the explosion in the burrow. Carbon monoxide, on the contrary, is poisonous to both animals and human beings if as much as one-tenth of one per cent is present in the air. This carbon monoxide is the principal, cause of the poisonous effects from illuminating gas in which it may be present in amounts of two to twenty per cent. The effect of the sulfur dioxide is less well known. The Selby Smelter Commission 6 found five-hundredth s of one per cent of sulfur dioxide caused a sense of acute suffocation to human beings, but no exact information has been found as to its effect on squirrels. EXPERIMENTS WITH SQUIRRELS In the earlier of these experiments an exact mixture of carbon bisulfide and air was made up in a series of communicating carboys. The outlet of these carboys was connected with the lower opening of a glass bell jar of sufficient size to contain a wire cage in which the squirrel was placed. A gentle stream of water was then passed into the farthest carboy which caused a steady current of gas to flow through the bell jar. In this manner the squirrel, kept in an atmos- phere of constant composition, was at all times under observation. Only a portion of the gas contained in the system was used in any set of experiments. The composition of the part which the squirrel breathed therefore could not be changed by coming in contact with the water. The plan of this apparatus is indicated in Figure 1. In this manner squirrels were treated with air containing 4%, 2%, and 1% carbon bisulfide. CONTROL OF GROUND SQUIRRELS BY THE FUMIGATION METHOD 211 i I— 4 2, O 1—4 3 o g C7i (W P3 O)-* 212 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION The detailed results of the treatments are given in Table I. It was found that 4% gas did not kill the squirrels more quickly than 2%, while 1% was too weak to be fatal within less than an hour and a half. With 2% gas the squirrels partially collapsed in from six to ten minutes and in seventeen to twenty minutes had ceased breathing. TABLE I. EXPERIMENTS WITH VARYING CONCENTRATIONS OF CARBON BISULPHIDE Percentage Manner of use of gas 4% Unexploded, constant current of gas 4% Unexploded, constant current of gas 4% Unexploded, constant current of gas 2% Unexploded, constant current of gas 2% Unexploded, constant current of gas 2% Unexploded, caged, artificial burrow 2% Unexploded, caged, artificial burrow 2% Unexploded, caged, artificial burrow 1% Unexploded, constant current of gas 1 % Unexploded, constant current of gas 4% Exploded, caged, artificial burrow 2% Exploded, caged, artificial burrow female pregnant 15 min. Sex Weight Time of Treatment Result female 22 min. death male 537 g. 13 min. recovered male 575 g. 14 min. recovered male 651 g. 21 min. death male 633 g. \l l A min. death female small 20 min. death male large 20 min. death female small 21 min. death male 537 g. 1 hr. recovered male 575 g. l^hrs. death female small 16 min. death 2% Exploded, caged, artificial burrow male old, large 20 min. 2% Exploded, caged, artificial burrow female small 23 min. 2% Exploded, caged, artificial burrow female large, strong 39 min. death death death death An artificial burrow was next constructed from 5-foot lengths of six-inch iron soil pipe. By joining several of these together with wet clay it was possible to modify the length and volume of the bur- row and also change its elevation when desired. A length of eleven feet was first employed. The upper end was closed by a tight wooden cover firmly wired on and plastered with clay to prevent escape of gas. In this wooden cover an inch hole was bored and closed with a rubber stopper. Through this hole measured amounts of carbon bisulfide could be added and on vaporizing could readily flow toward the other end of the pipe which was placed a few inches lower. The squirrel was placed in a large glass specimen jar, the upper end of which was covered with wire screen. The mouth of this jar was shoved into the expanded end of the pipe and the space between the neck of the jar and the pipe packed with clay The squirrel remained in plain view from the outside and could not escape into the pipe, but still received gas just as readily as if he had been within the pipe. The form of the artificial burrow is shown in Figure 2. CONTROL OF GROUND SQUIRRELS BY THE FUMIGATION METHOD 213 A number of squirrels were treated with carbon bisulfide in this artificial burrow and succumbed in approximately the same time with both 2% and 4% gas as did those which had been placed in a stream of gas in the glass bell- jar. A second group of squirrels were now treated in this artificial burrow with 4% and 2% gas which had been exploded. The carbon bisulfide was added through the hole in the cover at the upper end of the burrow and three minutes was allowed to elapse in order to give the liquid time to vaporize. A lighted match was then thrust in the hole and a vigorous explosion followed. The white smoke resulting from the explosion was driven through the pipe and into the glass jar containing the squirrel. It was evident that the explosion had some slight effect in distributing the gas more thoroughly throughout the pipe. None of the squirrels in any of the experiments appeared to be at all harmed by the explosion. They were clearly frightened but did not appear to be injured in any way by the concussion or by the flame which generally ran through the pipe. The squirrels which were treated with the exploded gas showed a little greater variation in the effect which the gas had upon them than had been the case with the unexploded carbon bisulfide. In general, however, they succumbed in practically the same time from the effects of both the exploded and unexploded gas. Autopsies were performed by Dr. G. H. Hart on a number of the squirrels which had succumbed to the exploded gas and on others killed by the unexploded mixture. The animals killed by the unex- ploded gas did not show any characteristic lesions. Their appearance was in all ways similar to that of an animal which had been killed by any anaesthetic such as ether or chloroform. The squirrels killed by the exploded gas exhibited the pink coloration of the lungs which brightens on exposure to the air and which is commonly caused by sulfur dioxide reacting with the coloring matter of the blood. There was no definite general bright coloration of the body cavity character- istic of carbon-monoxide poisoning, but samples of the blood examined by the delicate spectroscope test clearly showed the presence of carbon monoxide, indicating that this gas had contributed to the death of the squirrels. In order to determine what portion of the effect of the exploded gas was due to the sulfur dioxide, two squirrels were treated with a constant current of this gas in the glass bell-jar previously described. The first squirrel was kept in *4% sulfur dioxide gas which is five times as strong as the Selby Smelter Commission has found unsup- portable by human beings. At the close of an hour the squirrel was 214 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION entirely unharmed and did not appear greatly inconvenienced at any time. A second animal was then placed in 2% sulfur dioxide and after thirty-seven minutes exposure succumbed. It was not overcome by the gas at an early stage of the treatment, as is the case with carbon bisulfide, either exploded or unexploded. In fact, it was able to move about till just before its death and had it been removed earlier would probably have recovered. Sulfur dioxide is apparently about half as deadly as carbon bisulfide of the same concentration, but twice as much sulfur dioxide is formed by the explosion. Considering the sulfur dioxide alone we might expect that the exploded and unex- ploded gases would be equally poisonous, but if sulfur dioxide were the only poisonous substance present the squirrel would be overcome very slowly. Since the squirrel is overcome and collapses in prac- tically the same time with the exploded mixture, we infer that the small amount of carbon monoxide formed contributes materially to the effects causing the death of the animals. The only point which was noticed that indicated that the exploded gas might be any more effec- tive than the unexploded, was the action upon fleas. The fleas found upon the squirrels killed in the exploded carbon bisulfide were all dead, while those on the squirrels killed in the unexploded gas revived soon after they were brought into the open air. FLOW OF GAS UPHILL The next point studied with the artificial burrow was the rate of flow of gas around corners and up-hill. This is a matter of great importance in the efficiency of practical operations in the field. It is generally difficult to say which direction a burrow takes under- ground. It will therefore be impossible to tell whether the carbon bisulfide, applied with a waste ball, is placed at the highest point of a burrow or whether the farther runways rise to a higher elevation. If the gas cannot rise over elevations of one to two feet it is extremely doubtful whether the fumigation will be effective. The soil pipe was now extended to four lengths and two right angle bends placed in it. The bottle with the squirrel was inserted in the upper end of the pipe and the carbon bisulfide added through the opening in the lower end. The difference in elevation was slightly over eighteen inches and the distance between the two ends was twenty-two and a half feet. The first squirrel was fumigated with 2% carbon bisulfide. Ap- proximately thirty minutes elapsed before the squirrel appeared to detect any odor of the gas and a treatment of twenty minutes more was entirely without effect, the squirrel being apparently uninjured. CONTROL OF GROUND SQUIRRELS BY THE FUMIGATION METHOD 2\.) A second squirrel was treated with 2% carbon bisulfide and the gas exploded. The explosion was not vigorous, because the heavy vapor had not spread sufficiently far up the pipe. At the end of an hour the squirrel was still unhurt. The lowest point of the pipe was now raised about six inches so that the vajor could spread through this length and 2% gas was again exploded. The fumes of the explosion reached the squirrel but not in sufficient quantity to affect it seriously in an hour's treatment. The pipe was again lowered so that there was eighteen inches difference in level between the two ends. A squirrel was placed at the upper end and sufficient carbon bisulfide added to make a strength of 4% gas in the air. After half an hour's treatment this squirrel began to show signs of partial collapse but at the end of an hour could still move about briskly. It was taken out at this time, but died later in the night. Another animal was treated with 4% gas which was exploded and at the close of an hour was not seriously affected. A large strong squirrel was now placed in the upper end of the pipe and two ounces of carbon bisulfide was added to an inex- pensive pump machine which vaporized the carbon bisulfide and drove the gas with moderate force into the burrow. In ten minutes after the pumping began the squirrel commenced to be affected. Pumping was continued at infrequent intervals for ten minutes more. In thirty-five minutes from the time the treatment started the squrrel was dead. The conclusion to be drawn from these experiments is that carbon bisulfide gas will flow readily on the level or down-hill. It will not in either its exploded or unexploded form rise over an elevation of even one to two feet unless driven by some form of pumping or com- pressed-air apparatus. This undoubtedly explains many failures under field conditions. At the close of the carbon-bisulfide fumigations two squirrels, a large strong male and a female, were treated in the artificial burrow, with 2% Kilmol. The gas was allowed to flow from the upper end just as the 2% carbon bissulfide had and the squirrels died in twenty and twenty-one minutes, respectively, i.e., in the same time as did those treated with carbon bisulfide. A number of queries have been received as to the possibility of using other gases than carbon bisulfide for the purpose of exterminat- ing squirrels. These were the result of the accounts which had been published of the use of gas in modern warfare, though occasional attempts to try other fumigants have been made in the past. A great variety of gases are known to have been employed on the European battle front. The preparation of many of these is extremely com- 216 PNIVERSITY OF CALIFORNIA EXPERIMENT STATION TABLE II. EXPERIMENTS WITH OTHER GASES. Percentage Manner of use Sex Weight Time of Result of gas Treatment Approx. 4% Gasoline volatilized incom- gasoline pletely in carboys male 652 g. 43 min. recovered Approx. 4% Gasoline volatilized incom- gasoline pletely in carboys female 435 g. 52 min. death Approx. 2% Gasoline volatilized incom- gasoline pletely in carboys male 652 g. 1 hr. unharmed Approx. 2% More completely vaporized gasoline in simple carburetor male 652 g. 1 hr. *unharmed Approx. 4% More completely vaporized gasoline in simple carburetor male 652 g. 50 min. death Approx. 4% Exploded in artificial bur- gasoline row female large 25 min. unharmed 2% Kilmol Unexploded, caged, artificial burrow male large 20 min. death 2% Kilmol Unexploded, caged, artificial burrow female 405 g. 21 min. death Approx. 2% Caged, artificial burrow male large 42 min. death carbon tetra- chloride Chlorine Caged, artificial burrow female 624 g. 30 min. death Sulfur Undecomposed male 555 g. 2 hr. 15 min unhurt Chloride Sulfur Decomposed with water male 555 g. 3 hrs. partial Chloride recovery M% sulfur Constant stream of gas female large 1 hr. unhurt dioxide 2% sulfur Constant stream of gas female large 37 min. death dioxide Arsine-calcium arsenide 2 g. Small jar male large 25 min. death 15 g. Caged, artificial burrow male 701 g. 50 min. death 7 . 5 g. Caged, artificial burrow female 535 g. 45 min. death 7.5 g. Caged, artificial burrow female 470 g. 38 min. death 5.0 g. Caged, artificial burrow female 492 g. 46 min. death 7 . 5 g. Caged, artificial burrow female 414 g. 15 min. death after 2 hrs. 7.5 g. Caged, artificial burrow female 501 g. 10 min. death 55 min. later 7 . 5 g. Caged, artificial burrow male 164 g. 10 min. death after several hrs. (I day intervened after previous experiment) CONTROL OF GROUND SQUIRRELS BY THE FUMIGATION METHOD 217 plicated and expensive, and great danger would attend their use. Chlorine is probably the most readily available of those which can be employed. One great difficulty in regard to the use of chlorine is that the liquefied gas is at present put up only in large cylinders, in most cases containing 100 pounds, though a few eastern firms make 25-pound cylinders. The price recently quoted was 25 cents per pound, exclusive of the cost of the container. As a preliminary experiment one squirrel was treated with chlorine generated on a small scale in the laboratory. The effect of the gas was evidently extremely painful and irritating, very differ- ent from the carbon bisulfide which may be slightly exciting or intoxi- cating at first, but later acts like an anaesthetic. The squirrel also survived for half an hour, a distinctly longer time than with carbon bisulfide, and made vigorous efforts to escape. These efforts would doubtless result in thejreeing of many animals under field conditions. Chlorine is an extremely corrosive irritating gas. Its use would in- volve great danger to the operators as it readily attacks the mucous membrane of the throat and nose. The use of a gas under pressure from small cylinders would have several advantages as it would take the place of a pump machine, but it is not believed that this would be less expensive than carbon bisulfide. Gasoline has had a number of advocates and has on a number of occasions been used in local campaigns. A mixture was accordingly made up in the glass carboy system which would have represented 4% of gas in the air if it had been completely volatilized. One great difficulty was observed with both gasoline and distillate, in that a con- siderable portion of the material does not evaporate and is therefore wasted. One squirrel was treated with this approximately 4% gas till it partially collapsed, which required 43 minutes. It was then taken out and next morning was completely recovered. A second animal was treated with the same gas till it ceased to breathe. This required 52 minutes, and it failed to revive. An attempt was then made to devise a simple type of vaporizing carburetor, but even with this some gasoline remained as a liquid. A squirrel was treated in gas from this vaporizer and expired in 50 minutes. Two squirrels were then treated with approximately 2% gasoline vapor, one with vapor from the liquid evaporated spontane- ously in carboys and the other with vapor from the simple carburetor and in each case after an hour's exposure the squirrel was entirely unharmed. As a final test a mixture of approximately 4% gasoline vapor was made up in the artificial burrow, a squirred was placed at the lower end and the gas was exploded. The squirrel was greatly 218 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION frightened but was entirely unharmed either by the explosion or by the gaseous products formed. These experiments clearly show that gasoline is much less effective than carbon bisulfide and even with gasoline at 20 cents a gallon and carbon bisulfide at ninety, the carbon bisulfide would be the less expensive material. Carbon tetrachloride suggested itself because its vapor is heavy and non-inflammable. A squirrel treated with it died in 42 minutes. The effect was much the same as with carbon bisulfide except that it was much less rapid. It is also a more expensive material, so that these two considerations would rule it out. Sulfur dioxide, formed by burning sulfur has been used occasion- ally, but the results of the experiments heretofore reported do not warrant its application. Sulfur chloride, a heavy unpleasant liquid formed as a by-product in the manufacture of chlorine, has been found very poisonous and unpleasant for workmen around chemical plants. The liquid itself is partially volatile and is also decomposed by water into hydrochloric acid and sulfur dioxide. Two squirrels were treated with sulfur chloride, the first with the undecomposed liquid. After two hours it was entirely unhurt. The other was subjected to the fumes formed from the decomposition of the sulfur chloride and water and in three hours' time was only partially overcome. The material was abandoned as valueless. Dr. L. H. Duschak has suggested calcium arsenide as a possible fumigant. This substance when treated with water yields arsine, an extremely deadly gas. A sample containing 50% calcium arsenide was supplied by the Bureau of Mines and a series of squirrels treated with it in the artificial burrow. Amounts of five to fifteen grams, approximately one-sixth to one-half ounce, were used. These amounts corresponded to 3.5% to 10% of arsine in the air of the pipe. The calcium arsenide was placed in a small dish in the upper end of the burrow and an ounce of water was run into the dish through a glass tube. The decomposition of the calcium arsenide was practically instantaneous and the gas rapidly flowed to the lower end of the pipe. The arsine, unlike carbon bisulfide, is an active poison and not an anaesthetic. It does not, however, appear to attract the attention of the squirrel or cause any pain or discomfort until it produces a more or less active nausea. Where the squirrels were left in the gas con- tinuously they died in thirty-five to fifty minutes. This is not, of course, so rapid in its action as is carbon bisulfide, but it was found that if squirrels were left in the gas for from ten to fifteen minutes CONTROL OF GROUND SQUIRRELS BY THE FUMIGATION METHOD 219 and then removed to pure air, they still died in from one to two hours afterwards. This suggests that the calcium arsenide may possibly be very effective in the field, for even if a squirrel only came in con- tact with the gas for a brief period of time and then escaped it would be very likely to succumb later. The calcium arsenide is the only new material which appears to be especially promising, but it is not at present on the market and will require careful trial before it can be recommended for general use. FIELD FUMIGATION EXPERIMENTS At the beginning of the investigation it was found that very little exact information existed as to actual field conditions under which the fumigation has to be carried out. That is to say, nothing definite was known as to the size, depth, or extent of the burrows of the Cali- fornia ground squirrel. Various guesses were current that the depth of the burrow would range from two to six or more feet, and the area covered was equally indefinite. Shaw of the "Washington Agricultural Experiment Station has published a series of plans of the burrows of the Columbian ground squirrel, 7 and Frandsen of Nevada similar drawings of the Oregon ground squirrel, 8 but no such work had appeared on the California squirrel. A series of preliminary studies were accordingly made in the Berkeley hills in collaboration with Mr. Joseph Dixon of the Museum of Vertebrate Zoology, University of California. This work was later greatly extended by Dixon in the San Joaquin Valley, and observa- tions on the Oregon ground squirrel were made by Grinnell and Jacobsen in Siskiyou County. Detailed maps of portions of this work will be published elsewhere by Grinnel and Dixon, but the essential facts derived from it are given in Table III. TABLE III. CUBIC CONTENTS OF SQUIRREL BURROWS Length Diameter Sex Cu. Contents including chambers Result of Fumigation Berkeley Hills 5 ft. 3.5 in. male 1 . 03 cu. ft. killed squirrel 8 ft. 4.0 in. male 1.40 cu. ft. killed squirrel 14 ft. 4.5 in. male 2 . 40 cu. ft. killed squirrel 22 ft. 4.5 in. female 4.8 cu. ft. killed squirrel Bakersfield 34 ft. 4.5 in. male 4.8 cu. ft. killed squirrel (vicinity) 138 ft. 4.75 in. Colony 17.78 cu. ft. squirrel escaped Fresno Area 26 ft. 5.0 in. male 4.6 cu. ft. squirrel escaped Oregon Ground Squirrel Siskiyou County 66 ft. 2.5 in. female 3.48 cu. ft. no fumigation 220 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION The great variation in length and consequently in cubic content is immediately evident. In the Berkeley hills the male burrows were found on the outer edge of the ground occupied by large colonies. These burrows were in several cases quite short, though in others they extended to a length of fourteen or fifteen feet. The greatest depth reached by the male burrows in this region was eighteen inches to two feet, while the complete female burrow examined, reached a length of twenty-two feet and an extreme depth of thirty inches. Some of these burrows were treated by the waste ball method, both exploded and unexploded, and some by pumping. In all of them the animals were found killed near the nest. The ground was level or gently sloping and therefore presented no difficulties of treatment. Under such con- ditions any method of treatment by fumigation would be effective. The burrows treated by Dixon in the Bakersfield and Fresno areas of the San Joaquin Valley were in each case dosed by the waste ball method. The first one, in the vicinity of Bakersfield, was a compara- tively simple male burrow thirty-four feet long. One entrance to the burrow was closed and in digging the squirrel was found some twenty feet from the entrance. On digging further it was found that there was another entrance at the other end of the thirty-four feet. This had been unobserved and was left open, but the squirrel was neverthe- less killed. The ground was almost level and here again the gas had a perfect chance to flow. In the other burrow no squirrels were killed because it was a large colony burrow of greater length and with greater cubical air contents. The burrow dug out in the Fresno area was located in the "hog wallow" country and is an excellent illustration of the extremely difficult conditions which may be encountered. The land has many layers of hardpan which are not continuous and between these the squirrel's burrow twists and turns. The side view of this burrow is reproduced here (figure 3) as it shows so clearly the problem involved. The burrow in question was treated by the waste ball method. It was then dug out and the discovery was made that the gas had not risen over "c"; the squirrel being found alive in the nest. This points to the necessity of using some form of apparatus to pump the gas. The depths of the burrows in the valley areas were found to be greater than those in the Berkeley hills where the shale subsoil tends to limit the habited zone. The greatest depth reached in the San Joaquin Valley was four to five feet, which is believed to be typical of conditions in the interior valleys. The Oregon ground squirrel is only of importance in the extreme northern section of the state, but where found is exceptionally numer- CONTROL OF GROUND SQUIRRELS BY THE FUMIGATION METHOD 221 o ^ c Ui P H O „ 05 5- Hs M . N -' P 5T ►-• b ^ g P CD - 3. 5" ° o -cr o r+ CD 2 S x p ^ ^ o °-i <*- o B o Hop O