WOOLLY APHIDS OF THE ELM. 2O9 



Fig. 134. The antenna of this generation, is 6- jointed and 

 distinguished from that of the rosette dweller in having III 

 typically shorter than IV+V-f VI. (32A-13.) This generation 

 matures about June 17 in Maine. 



Whether the migrant generation maturing late in June are 

 altogether the progeny of the second apterous generation or in 

 part the daughters of the stem mother I do not know. My field 

 observations lead me to think that part of the progeny of the 

 second apterous generation desert the elm leaves and settle 

 under the scales of the elm bark in tender places on the trunk 

 and there subsist as bark feeders in woolly colonies but less 

 conspicuous and more hidden than rileyi. At any rate such bark 

 colonies occur here in Maine for which I can not otherwise 

 account. 



The fact that the stem mother is still actively giving birth to 

 nymphs for some time after the second apterous generation 

 become mature and are producing young complicates the case 

 considerably when an attempt is made to sort out the young. 

 Figure 139B shows an old roll just deserted by the migrants. 



Fig. 135. The antenna of the spring migrant. It is typical 

 for this structure to have III shorter than IV-j-V-j-VI and VI 

 subequal to or even longer than V. The length of VI, however, 

 is variable in Maine material and is apparently ordinarily a little 

 shorter than with the same species as it occurs in the western 

 states. The number of annulations on III is also subject to con- 

 siderable variation. The antennae of those individuals develop- 

 ing in leaf curl late after the leaf becomes less sappy are likely 

 to differ very strikingly from those developing earlier in the 

 season both in total length of antenna and in the number and 

 distribution of the annulations, as is illustrated in Annals Ento- 

 mological Society of America, Vol. 6, Plate 24. Figs. 2, 6 and 

 7 are all drawn to the same scale. Figs. 6 and 7 represent the 

 antenna of individuals collected July 23, 1912 for comparison 

 with specimens, Fig. 2 which developed June 30 in the same leaf 

 rolls. The difference in the tola'l size of the antennee and in the 



