BEHAVIOUR OY APOSEMATIC SPECIES 211 



out the legs so as to display the bright pink (a typical 

 aposeme), and emitted from the thorax just behind the 

 head a copious yellow froth, till a mass of yellow bubbles, 

 with a very strong acrid odour and taste, projected on 

 each side." 



Equally remarkable is another specialized grasshopper, 

 known to the Baganda as " Semukutu." ^ It is a large 

 heavy insect of dull greenish-grey colour, and though 

 not in any way aposematic in colouration, yet has all 

 the characteristics of such an insect. It has no trace 

 of wings or wing covers, and the thorax is set with spines 

 around its margin. The abdomen is fat and bloated, 

 and the sluggish insect makes no attempt to evade an 

 approaching enemy. I could never induce a monkey to 

 tackle one of them. When handled, a Semukutu ejects 

 with some force from the sides of the thorax a yellow 

 fluid of strongly acrid odour, which would prove extremely 

 unpleasant in the mouth ; and, indeed, the Baganda will 

 not even handle the insect, for they say the fluid produces 

 sores on the skin.^ But I never found that it had any 

 effect on my skin, and I have handled a great many 

 Semukutu. This is an interesting case, for here we have 

 an insect which has acquired the habits and attributes 

 of an aposematic species, yet not the colouration. Indeed, 

 a young specimen is rather procryptic, being mottled with 

 green, but it has not then the power of exuding the 

 acrid fluid. One might, however, suggest that the appear- 

 ance of this stout, spiny, wingless grasshopper is so unlike 

 that of others that it can be very easily recognized, and 

 the need for aposematic colours may not have arisen. 



The above observations on the monkey and grasshoppers 

 have exemplified the deliberate use of an aposeme by an 

 insect when hard pressed. This is again brought out 

 very well by an experiment with fowls, the subject being 



^ Enyaliopsis, possibly durandi. 



* See Bulletin of the Entomological Research Committee, vol. i, p. 227, 



