III. Inlieritance in Shirley Poppn 
411 
stamen. Further, we have at present no measure of the homotyposis of plants 
in relation to supernumerary petals and petaloid stamens. 
Colour of Middle Third of Petal. This was worked by a fourfold table, the 
division being into poppies showing some and tliose showing no red. We cannot 
do more than consider the maternal plants as an average random selection of 
Shirleys (see p. 406). In the matter of body colour, within the classifications red, 
pink, pink-white, etc., there appears to be small variation within the plant, or we 
may look upon this broad colour classification as individual to the plant and not 
to the bud. 
The Basal Patch. This was dealt with by a fourfold table, the division 
being, none to slight, and slight-definite to large. No evidence is available 
as to whether there was or was not a selection of maternal plants. The extent 
of basal patch varies considerably on flowers of the same plant, and is probably 
a good deal influenced by personal equation and homotyposis. 
Wrinkling. The division here was by a fourfold table. Mr Blanchard's esti- 
mates of wrinkling were so low as compared with Mr Ward's that the division 
of the maternal plants was made into not wrinkled and some wrinkling, wliile 
the Crewe population was divided into wrinkled and slight or no wrinkling. 
There may be a selection stringency of "8 to '9, and probably normal homotyposis, 
as wrinkling varies considerably on the same plant. But we have no definite 
knowledge on these points, and we consider wrinkling, although very characteristic 
of the Shirley, as very unsuitable for record. 
We sum up our results in the following table : 
TABLE III. 
Parental Correlation, Parkstone Maternal and Crewe Daughter Plants. 
Character 
Kaw 
Correlation 
Selection 
Stringency 
Corrected 
Correlation 
Stigmata... 
Extent of Margin 
Number of Petals 
Colour of Middle Third 
Basal Patch 
Wrinkling 
•1717 
•3175 
•2740 
•4008 
•1305 
•1595 
•7295 
•6259 
•5278 (?) 
0 
•6900 (??) 
•8620 (??) 
•2324 
•4718 
•4750 (?) 
•4008 
•1864 (?) 
•1842 (?) 
Mean 
•24 
•69 
•33 
We have not the least hesitation in admitting the full inconclusiveness of 
these results. All we consider that they show is, that if we could accurately 
allow for selection of parentage, for the variability of character in a plant with 
its multiplicity of organs, and for the extreme response to seasonal and environ- 
mental changes, and, further, eliminate the effects of personal equation, there is 
no reason to suppose that the parental resemblances for different characters would 
52—2 
