518 



SCIENCE. 



[N. S. Vol. XII. No. 301. 



pregnates the atmosphere and frequently 

 causes a smarting sensation in the eyes and 

 nose, even when the animals are held at 

 arm's length. At close range the eyes 

 would certainly be most painfully if not ' 

 seriously affected, and such injury is prob- 

 ably the ground for the fear in which large 

 diplopods are held by the natives of many 

 tropical countries, although the creatures 

 are otherwise quite harmless and may be 

 handled with impunity. In Porto Eico, 

 for instance, it is universally believed that 

 the local Spiroholus has a deadly sting in its 

 tail. 



When the repugnatorial liquid comes in 

 contact with the skin of the hands a yel- 

 lowish-green stain results, which gradually 

 deepens to a dull purple. In one instance 

 after collecting a considerable number of 

 large iSpirostrepti at Conakry, Senegambia, 

 my hands, although washed soon after, be- 

 came very deeply colored and the skin 

 peeled from the stained areas a few days 

 afterward. As is well known to students 

 of the group, the alcohol in which these 

 large diplopods are collected takes on similar 

 colors, yellowish green at first, changing to 

 a very deep purplish red, and has a char- 

 acteristic disagreeable odor different from 

 that of the living animal, but still in some 

 respects suggesting it. This odor Dr. Loew 

 considered similar to that of pyridine, 

 though the presence of that substance 

 might possibly be due to a taint of ' denat- 

 uralized ' German alcohol and would not 

 explain the corrosive action of the fresh 

 fluid upon the cells of the epidermis, which 

 fact denotes a peculiarity of chemical ac- 

 tion known in very few organic compounds, 

 and possessed neither by prussic nor by 

 formic acid, now recognized as the poison 

 of the centipedes as of many other ani- 

 mals. 



The volatile character of the fluid, or at 

 least of its active constituents, is also as 

 apparent in Sjnrostreptus and Spirobohis as 



in Polyzonium and Polydesmus. Until dis- 

 turbed and roused to defensive efforts the 

 creatures are nearly or quite odorless, and 

 some will bear considerable handling before 

 discharging their batteries. Indeed that 

 operation can scarcely be supposed to give 

 pleasure to the animal since its own fumes 

 are soon fatal when it is confined with them 

 in a bottle or small box. The same is also 

 true of Polydesmus, the ability to secrete Prus- 

 sic acid involving no immunity to that poison 

 when the creature is obliged to breathe its 

 own gaseous secretion in a closed space. 

 The possibility that a certain amount of re- 

 sistance to prussic acid may, however, be 

 developed in other animals is suggested by 

 reliable testimony to the effect that one of 

 the monkeys native in Liberia is fond of 

 the locally common species of Oxydesmus 

 (fiavoniarginatus. medius and grayi). There 

 was no opportunity for me to verify this 

 information, but places where the leaves 

 and vegetable debris had been upturned, 

 declaredly by the monkey in searching for 

 his prey, were pointed out to me in the for- 

 est, and the flesh of the animal is commonly 

 believed to have the odor of Oxydesmus and 

 to be bitter, poisonous and inedible because 

 of this remarkable and unique food habit, 

 on which point the observation and opin- 

 ion of the natives might be expected to 

 be conclusive, nothing being refused by 

 them which could by any possibility be 

 eaten. 



Direct exposure to the light and heat of 

 the sun is also speedily fatal to many Dip- 

 lopoda. In the case of small and delicate 

 species this might not seem suprising, but 

 the very large and heavily armored Sjnroboli 

 and Spirosirepti are also unable to recover 

 after full illumination for a short time. In 

 ten or fifteen minutes they are often quite 

 dead. That this susceptibility may prove 

 to be the result of some chemical change 

 or dissociation of the stored repugnatorial 

 fluid is apparently indicated by the fact that 



