EXPERIMENT STATION BULLETINS. 483 



death) of the actions of 45 of llie insects treated with borax, sodium 

 fluoride, or borax and p.vrethruni mixed. A few tests were made by 

 compulsory feedino' of the roaches with a very little borax or sodium 

 fluoride from the point of a small knife-blade. In every case the symp- 

 toms were much like those described above and the insects finally died. 

 These last tests made it clear, therefore, that both borax and sodium 

 fluoride could act as stomach poisons in the case of the cockroach. 

 Moreover, it seemed very likely that in cleaning either of the powdered 

 poisons from antennae, feet and body, the insect would eat enough of 

 the poison to kill. Finally, visible proof was obtained that powder, 

 licked from various parts of the body, was swallowed. Sodium carmine 

 was very thoroughly dried and pulverized with borax, and with sodium 

 fluoride. Insects were placed with this stained powder, and several 

 hours after they had cleaned it from their bodies, dissections were made; 

 the red stain w^as found in the contents of the crop. 



The question still remained, however, as to whether a little of the 

 powdered insecticides might not be dissolved in the moist exudations at 

 the bases of the legs and betAveen the sternal plates, and then absorbed 

 through the thinner body integument in the locations named. It seemed 

 possible also that the very fine powder might enter the spiracles. That 

 is, it was still a question as to whether the powders under consideration 

 might not be contact insecticides as well as stomach poisons when used 

 against roaches. In order to get at the solution of this question, it was 

 found to be absolutely necessary to devise some way of preventing the 

 insects from licking their bodies after they had become dusted with the 

 powder — without, at the same time, inflicting any serious mechanical 

 injury in the use of the yjreventative device. The desired result was 

 accomplished by means of card board collars, which were prepared and 

 Ksed in the following manner: 



The collars were made from single pieces of flat bristle board of about 

 ihe weight of ordinary medium index cards. Each collar Avas four to 

 four and one-half inches in diameter — i. e. large enough to form a 

 cover over a shallow glass stender dish. A slit Avas cut in the circular 

 card from the circumference to the 'tenter, Avhere a small smooth hole 

 was punched. This central opening Avas large enough to turn freely on 

 the neck of a cockroach, but it Avas very much smaller than the animal's 

 head. After the collar ha 1 been very carefully placed around the neck 

 of a roach, a thin strip of jiajer was afterward pasted over the slit. 

 The flat collar could then be placed over the stender dish, containing a 

 liitle powder to be tested, so that the insect's body rested upon the bot- 

 unn of the dish, and the head stuck up above the collar outside the dish. 

 The fluttering Avings and scrambling legs would thoroughly dust the 

 animal's body without any of the powder's reaching the head or mouth. 

 Fig. No. 4 of Plate I .shoAvs flic i)0.silion of tlie head of a cockroach Avear- 

 ing such a collar. In check tests, untreated roaches Avere made to wear 

 collars for eight, eighleeu, and even ninety-eight hours Avithout any 

 serious injury. After these specimens Avere released, they were kept 

 under observation for ten to nineteen days, and during that time fed and 

 behaved normally. In all these experiments the roaches Avere permitted 

 to have moist food Avhen no insecticide poAvder Avas present and they 

 Avere not wearing a collar. 



