W. F. R. Weldon 247 
with its seed-coat uninjured. A statement by Culverwell seems however worth 
quoting. In 1882 he writes of Telegraph and Telephone : " The two will always 
" come from one sort, more especially from the green variety." And a line or two 
later, describing the result of sowing what he thought a good sample of Telegraph, 
he says, " Strange to say, although the peas were taken from one lot, those sown 
"in January produced a great proportion of the light variety known as Telephone. 
" These were of every shade of light green up to white, and could have been 
"shown for either variety." {Gardeners Chronicle, July 1882, p. 150.) 
This is only one of a number of statements, scattered through the Gardener's 
Chronicle from 1879 onwards, which show that neither Telegraph nor Telephone 
was recognised as sharply divided into green or blue and yellow or white races 
during the early years of their existence ; Culverwell's statement that the green 
variety was especially variable in colour in 1882, shows more than this, however, 
because it shows that in his opinion a plant of a late generation, which exhibited 
a "recessive" colour, did not produce exclusively recessive offspring, as Mendel 
says it should. At the present day they are recognisably distinct, both in shape 
and in cotyledon colour ; but the colour of the cotyledons is variable in both. 
Using the groups of seeds photographed on Plate I. as a scale, I found among 
576 seeds of Telegraph, 512 seeds of fairly uniform colour, and 64 obviously pie- 
bald seeds. The self-coloured seeds fell into one or other of the colour groups 
from 1 to 6, with frequencies given in Table IV. (p. 250). The piebald remainder 
were generally half bright yellow and half bright green. There were generally 
only two patches of colour on each seed, and one cotyledon was often as yellow 
as Group 6, Plate I., while the other was as green as Group 1. 
(2) Pride of the Market. If shape and colour be considered together, this 
is one of the most constant races selected from the offspring of Telegraph. Of 
500 seeds whose shape was examined, there were 
Like Fig. 1 260 
„ Fig. 2 218 
„ Fig. 3 20 or 21 
„ Fig. 4 2 or 1 
while the colour, as shown in Table IV., is distinctly less variable than in 
Telegraph, and the percentage of piebalds perhaps significantly smaller. 
(3) Stratagem is nearly as constant as Pride of the Market in colour, but 
it is more variable in shape. The seeds are larger than those of any other 
race belonging to this group, and are normally much flattened ; for this reason 
it is difficult to compare the seeds of Stratagem with those of the races just 
described. Three groups of Stratagem seeds are photographed in Figs. 7, 8 and 9 
('Plate II.) and they show the characteristic shapes and wrinkles fairly well. 
An inspection of Figs. 1 — 9 will show better than any words, first how completely 
the categories " smooth " and " wrinkled " pass into each other in these races ; 
