1U11] Specific Characters in Pseudococcus 313 



Antennae. — The number of segments of the antennae 

 is a generic and not a specific character, and so is of no import- 

 ance in separating species. Several species have been described 

 with seven segmented antennae. The number of segments, 

 eight, of the type of the genus, longispiniis Targ, has been 

 adopted in this paper. However, a dimorphism has been 

 described in two instances, viz. : by Folsom for trifolii Forbes, 

 and by Essig for agrifoliae Essig. In these descriptions there 

 is said to be a winter form which has seven segmented antennae ; 

 this form giving rise to the summer form with the normal eight 

 segmented antennae. If further investigation shows these 

 observations to be correct, it will be of both specific and generic 

 importance. 



The comparative length of the different segments of the 

 antennae is a character that is the most used of any of the 

 characters of the insect's body. The relative length of the 

 segments and the formulae deduced from the measurements is 

 almost invariably contained in descriptions. Sometimes a 

 considerable variation is noted, several formulae being given. 



In this study ten specimens of each species were used. 

 Aside from the question of variation which will be taken up 

 later, the relative length of the antennal segments is not a desir- 

 able character to use. The greatest difficulty in its use is the 

 difficulty of making correct measurements and the determina- 

 tion of the exact limits of the various segments. The chitin is 

 not continuous from one segment to the next and consequently 

 the portion between the chitinous parts of the segments, the 

 conjunctiva, is not visible or only slightly so in well cleared 

 specimens. Consequently, in making measurements, the deter- 

 mination of the end of a segment will be only approximately at 

 the center of the conjunctiva. This difficulty will be much 

 increased if there are some bends in the antennae. A second 

 difficulty is to determine some point at the end of the segment 

 from which the measurement will always be taken. This 

 difficulty is most apparent with the first segment. This seg- 

 ment is an irregular truncated cone with the sides of different 

 lengths, and but little longer than the width across the base. 

 It will be seen that the determination of the same points for 

 the measurement of this segment would be nearly impossible. 

 In the tables given the writer does not feel that the measure- 

 ments of the first segment are dependable. Another difficulty 

 in getting dependable formulae is the very slight difference in 



