EXPLANATION OF FORM AND COLOUR[NG. 331 



at once resumed its former shape each time the pressure was relaxed, the bug 



remaining absolutely motionless with walking-legs all stretched stiffly back 



under the protection of the abdomen, and claspers stretched stiffly forward but 



ready to suddenly cla.sp at whatever came within reach. By thus unexpectedly 



grasping the wires of the cage the bug was several times wrenched right out of 



the roller's bill and the latter, heartily annoyed, apparently, by this leathery 



object that kept catching in everything, at last flung it away in a rage and refused 



obstinately to take it again. She then seized a smallish Coprid with hard I'ounded 



glossy elytra., Copris orion $ ,but the beetle at once slipped out of her grasp. She 



refused to accept it a second time or an Amicmttcs glohulipennis, crushed well and 



threw right away the Lycoid bug, accepted a dark gi'een bug {Cryptacrtis comes 



var. apiccdis, Dist.), which slipped out of lier grip the first time (it is harder than 



Cryptacrus comes), but was easily enougli crushed on being reofFered and eaten with 



apparent relish in spite of its strong smell, which it was obvious was noted by the 



bird. She then went on once more to crush and reject the Bostrychid and to 



refuse without tasting the Cy^natura and the water-bug, the Coprid, a P. angolamis, 



and the P. demodocus, but to accept and eat most readily a P. cehrene. 



[Order : — 



1. Polyhiroiia cenigma {regretted"^), Cryptacrus comes 



var. apiccdis, Precis cehrene. 



Bostrichoplitescornutus, \ 



Lycoid bug, Red 

 Millipede, Laccotrephes 



2. Papilio angolanus, P. demodocus (on manner). 



ater, Copris orion $ and [3. Grasshoppers, Cymatura hifasciata (on manner), 

 probably Amiantus I 

 glohidipennis. ) 



P. angolanits and perhaps the Cymatura were eaten in preference to the leopard's 

 meat before this had yet disagreed. With its evident disagreement may be 

 compared the fact that it was only eaten by my carnivorous animals under stress 

 of hunger. 



The Polyhh'ma'a special defence had probably been much reduced by capture ; 

 the Lycoid bug was rejected more emphatically than the Bostrychid ; the special 

 defence of the Laccotrephes and the Copris are worth noticing, as also the fact 

 that Cryptacrus comes var. apicalis is harder and so distinctly better protected 

 against enemies in the manner shown by the experiment than typical Cryptacrus 

 comes. I have always regarded them as distinct species.] 



Exp. 130. — December 29. Persistently refused even to taste a Laccotrephes ater, 

 but readily ate a P. angolanus, a Mycalesis campina, and a T. hrigitta. 



[This places the water-bug relatively to the roller below P. angolanus, if not below 

 T. hrigitta, a,pparei:itly mainly if not entirely in vii'tue of its leathery covering.] 



Exp. 131. — January 7, 1910. To place HerjKenia erlphia and Teracolus phlegyas. 

 After several gi'asshoppers, readily ate a Mycalesis campina, tyvo large grasshoppers, 

 another Mycalesis, a P. angolanus, and a Teracolus phlegyas, yet another grass- 

 hopper and three more Mycalesis. The next grasshopper was held for some time 



LINN, JOURN,— ZOOLOGY, VOL, XXXIII, ^4 



