DEVELOl'MENT OF THE ANTENX.E OF TERMITES. 



279- 



Trinerviterines ai^e either XYII-, XYI- oi' X\'-joiiited. It 

 has never happened in my experience that all the winged 

 individuals from a nest have antenna? of a i-egular numl)er 

 of joints. Those of trinerviformis are generally of XA'II, 

 but may be of XYI joints. Those of a species taken to be 

 gemellus Sjustedt are generally XY, but may be X\'I. 

 Those of three undescribed species before mo may be XY,. 

 XYI or XYII, and the common number varies with the 

 l)roods of different nests. 



Text-fig. 10. 





Ixs^^^srsz g iK * 2J ^tt^ 



Eutermes (Trinervitermes) trinerviformis Holmgren. a-e. 

 Imago. /. Larva of the third instar. y, h. Larva ah ovo. i-n. Nasxitn. 



The homologies of these variously composed antennae is 

 illustrated by fig. 10, a-e. The XYII is reduced to a XYI 

 either by the union (13 + 12) as in fig. 10, h, or (III + 14) 

 as in fig. 10, c. This again is reduced to a XY when both 

 unions (13 + 12) and (III + 14) occur in the same organ as 

 in fig. 10, d, e. There is seldom any trouble in establishing- 

 the coinpound nature of the conventional III or lY, as the 

 case may be, as where they may appear simple, if lai-ge, as 

 in fig. 10, e, another organ will show them pseudo-articulated 

 as in fig. 10, d. 



The development of these antennas shows that in all 

 patterns the flagellum is one of 19 elements. The XYII 



