400 
Coojierative fnrestigations on Plants 
The first number (144) in the first column refers to the original Hampden 
great-grandmaternal plant, the second (30'16) to the Kidderminster I. grand- 
maternal capsule, the third (S26) to the Parkstone maternal capsule, and the last 
number is the key number of the Crewe crop. C. F. signifies that it was from a 
free bud, or 826 capsule was in all probability cross-fertilised. 
Again : 
Petals 
Base 
Pedigree 
('') 
Bands 
Colour 
Margin 
Wrinkling 
(«) 
Eemarks 
(«) 
149 
6 
3 
13 
r 
/; 
n 
dl 
w 
Conical topped capsule 
26-15 
D. 471 
This gives a grand-daughter capsule from ArnclifFe of an original Hampden 
plant. It will be seen that with the crop-books before one it is possible to rapidly 
look up the pedigree of any plant, although the pollen ancestry must remain 
unknown. 
By observing characters on the apical capsules only we at first considered that 
we should get rid of the difficulty of having to allow for homotyposis, i.e. the 
variability of character within the individual plant. This variability, as was 
pointed out in the first memoir, is very considerable. But as the work went on 
the question was again raised as to whether the apical flower is as true a measure 
of individuality as the totality of flowers on the plant. This point has been to 
some extent discussed in the first memoir on the Shirley*, and is by no means 
easy to answer. We found in that paper that offspring from seed of all the 
capsules of a plant had a greater average resemblance to the plant than offspring 
from seed taken from a single capsule, while the offspring were more like each 
other when their seed was taken from a single capsule. If the arrangement of 
buds on a plant were as definite as the digits of a man, we should have no hesita- 
tion in comparing the characters of the apical flowers of mother and daughter 
plants. But this is certainly not the case, and when we deal with characters 
which are by no means peculiar to the apical flower it becomes a very difficult 
point to determine how far the variation within the individual plant can or must 
be allowed for. Characters like " wrinkling," number of petals and number of 
stigmata have a very considerable variation within the individual. Frequently 
when the general tendency of the flowers of the plant is to a high number of 
stigmata, or to an abnormal number of petals, this tendency may fail to be repre- 
sented on the first flower. Other characters like colour and margin may possibly 
vary less from flower to flower of the same plant ; but are nevertheless sensibly 
influenced by environment. Lastly, a very considerable selection takes place of 
the cajDSules which give any germinating seed. The mother plants which provide 
germinating seed are far from a random sample of the general Shirley population. 
* Bio)iu'tril;((, Vol. ii. pp. 7.^, 74, and p. 82. 
