PAPERS ON ZOOLOGY AND PHYSIOLOGY 307 
Parent-Progeny Correlation in a Subline where Selection 
was Made in an Attempt to Change the Ratio of the Length 
of the Third to the Fourth Segments of the Antennae. 
The first character which was used in the selection work 
with Aphis avenae Fab. was the ratio existing between the 
length of the third and fourth segments of the antennae. Eight 
selections were made in attempting to increase this ratio in 
the fifteen generations obtained. Already there has been pub- 
lished (Ewing, 1914, b) a statistical study of regression in 
this subline or isolation, and here I will give the results of 
working out the parent-progeny correlation and add a few 
notes on regression. If we group the parents into classes each 
with a mean difference of 0.10 points in the antennal ratio, 
the length of the shorter segment (segment four) being taken 
in each case as unity, we will have the arrangement found in 
Table I. This has been done and the offspring are grouped 
into the same classes as the parents, thus making a symmetrical 
table. 
TABLE I. 
PARENT-PROGENY CORRELATION TABLE WITH RESPECT TO ANTENNAL FORMULA 
(RATIO OF LENGTH OF THIRD TO FOURTH SEGMENT.) 
Parents 
1.20-1.29 
Offspring 
1.20-1.29 
1.30-1.39 | es ae (ae A 
1.40-1.49 fis Fa] heel Oe ee 
1.50-1.59 
1.60-1.69 | 
1 
1/ 3] 2 | 4| 
1.70-1.79 1] 6| 5 [a1] 2 | 2| [24 
1.80-1.89 EE ee | 1) Bt 
3] 3 
2.10-2.19 Bia 
| Est | [2 
2.20-2.29 | 0 i es ee 
2.30-2.39 le PRE 26 Se a a a a | 
2.40-2.49 PE FG SRT ad 1 Cd a Peay 
2.50-2.59 ay — prenit! RE 
2.60-2.69 
2.70-2.79 
280-289 0) A Ae a eG YT 
hie | 
2.90-2.99 
An examination of this table shows a lack of correlation. 
The offspring of the three parent classes having the highest 
