298 Mr. Poulton's/^/'^/icT notes upon the 



shagreen dots. The annulatioii of the segments is 

 very regular and distinct, and the shagreen dots are 

 arranged in rings, one to each annulation (this is 

 always the case in shagreened larvae). The prothorax 

 has three or four such rings, but they are rather con- 

 fused ; the mesothorax has five, the metathorax six, the 

 1st abdominal seven, and the succeeding six segments 

 eight rings each. Posterior to these there are no rings 

 or annuli, with the exception of three rings upon some 

 irregular folds behind the caudal horn, which represent 

 (I believe) the 9th abdominal segment. 



0. Notes upon the adult larva of Macroglossa stella- 

 TARUM. — This ontogeny has been given very fully by Weis- 

 mann. I received a large number of larvae in the last 

 stage from Malvern during the past summer. The bifid 

 termination of the horn could be traced in several 

 individuals, even at this late stage. It is probably very 

 distinct at earlier periods. Tlie spiracle upon the pro- 

 thorax is red, while the eight others are black and far 

 more distinct. There is a black line round the former, 

 which can only be seen by careful searching. I have 

 not before noticed an instance of different coloration 

 among the spiracles. The segments are distinctly annu- 

 lated, except the 8th abdominal, and the number of 

 rings seems to be as in the larvae described (eight upon 

 most of the segments). In a rather contracted attitude 

 the segments taper rapidly from the 1st abdominal to 

 to the small head. When startled the head and thoracic 

 segments are sharply retracted into the 1st and 2nd 

 abdominal, which become swollen. (This is especially 

 true of the former segment.) When the larva is 

 stretched the tapering is not so distinct, and extends 

 further back. Thus we have a behaviour identical with 

 that of Chcerocanipa elpenor without the modification of 

 marking which have appeared in the latter. Weismann 

 and Meldola give instances of such terrifying attitudes 

 without the co-operation of terrifying markings, but it is 

 probable that the object of JM. steWitanim is simply that of 

 a rapid withdrawal from danger, and perhaps a partial 

 protection of the head. It is thus likely that the first 

 stage in the evolution of the terrifying attitude (such as 

 that of C. elpenor, &c.), is a case of protection in one of 

 its simplest forms. The habit becoming fixed and 

 producing further structural modifications as the 



