146 



largest about four and a half inches. The hind femora on 

 the small males have armature not much more pronounced 

 than on the females, but on the large males the two large 

 teeth are very strong, and there are up to twelve smaller ones ; 

 the femur itself on the large males is considerablv wider than 

 the abdomen, and passes the sixth segment of the latter: 

 whereas on the small males it is narrower than the abdomen, 

 and may not extend to the tip of the fourth segment. 



Six specimens (all males) were taken, on which, a leg 

 having been lost, it had been renewed; four of these were 

 larvae, and the renewed legs are very small (pi. xiii., 

 figs. 5-8). On the two mature specimens (pi. xiv.) the 

 renewed leg is a hind one, and it is remarkable that with 

 each the leg, in addition to being much smaller than the 

 non-renewed one, is without conspicuous masculine armature, 

 approaching the female type. 



I saw no communities in dead trees, but was told that 

 occasionally when a tree touched a house, and especially in 

 the olden days when many houses were thatched, the phasmas 

 would frequently take shelter in large numbers in the ceiling 

 and make uncanny noises at night. But they were easily 

 checked from doing this by preventing any part of the 

 adjacent trees from touching the house. In nearly all cases 

 that I examined it was seen plainly that the phasmas had 

 selected for their domicile trees that had been largely bored 

 by larvae of a large longicorn beetle. 



This beetle, Agrianome spinicollis, W. S. Macl.,^ is 

 abundant on the island, and numerous specimens may be 

 taken at night on the trunks of many kinds of trees. The 

 larvae make large borings and eject most of their excrement 

 outside the trees; each makes a large hollow, and as many 

 usually work together their borings frequently interlace; in 

 time, with the work of other insects, timber-rotting fungi, 

 etc., large hollows are formed, and in these the phasmas take 

 up their residence. A large orange tree was practically 

 destroyed by larvae of this beetle, and figures (pi. xvii.) are 

 given of the inside and outside appearance of portion of the 

 trunk of this tree, showing the great damage done. Large 

 Kent id trees are also attacked. No living banyan trees were 

 noticed to be attacked by it, but larvae were abundant in 

 dead ones. Very old and rotten logs of many kinds of trees 

 appeared to be attacked with as much favour as living trees, 

 and their larvae are often taken from such logs by the 

 islanders, with whom they are very favourite fish-bait. 



(l) It is not Cnemoplites (Arimaspes) howei, as the late Mr. 

 A. S. Olliff thought possible. 



