240 MendeTs Laws of Alternative Inheritance in Peas 
second generation invariably resemble those of P. arvense in shape, although in 
colour they follow Mendel's law of segregation ! 
The discussion of seed-coat colour is more difficult than that of the two 
characters already dealt with, because the recent writers (especially Tschermak) 
pay attention to a character, namely green pigment in the seed-coat, which 
Mendel does not mention. It is clearly unreasonable to expect that every 
character in the seed-coat should obey the same law, and therefore in trying to 
estimate the agreement between Mendel and later workers, many valuable data 
collected by Tschermak will be neglected. The most sticking exception to the law 
of dominance is that observed by Correns. The grime spate Erfurter Folger- 
Erhse has a nearly colourless seed-coat (recessive, Mendel) ; the P urpurviolett- 
schottige Kneifel-Erhse and the Pald-Erhse mit purpurrothen Hillsen have the 
seed-coat uniformly orange, becoming brown with age. In hybrids of the first 
generation, between either of these last varieties and the Erfurter Folger-Erbse, 
the coats of seeds (often in the same pod) were sometimes nearly colourless, some- 
times intensely orange-red, but generally more or less orange-red, and spotted 
more or less strongly with blackish violet. The seeds of extreme colour, those 
with orange or those with almost colourless seed-coats, gave plants of the second 
generation which again showed the same extremes of colour in the seed-coats, 
connected by transitional forms. This is clearly a case in which one of Mendel's 
characters obeys neither the law of dominance nor the law of segregation. 
The appearance of blackish-violet dots on the seed-coats of similar hybrids 
was noticed by Mendel. In his crosses however one parent showed traces of such 
dots, which were only exaggerated in the offspring. In the plants used by Correns 
as parents, the seeds with brown coats are not said to have possessed these dots. 
An exaggeration of the purple dots, of a kind similar to that observed by Mendel, 
but even greater in degree, was found by Tschermak in crosses between Graue 
Riesen and varieties of P. sativum with colourless seed-coats. 
Enough has been said to show the grave discrepancy between the evidence 
afforded by Mendel's own experiments, and that obtained by other observers, 
equally competent and trustworthy. It does not seem to me reasonable to doubt 
the substantial accuracy of any of the statements made by the observers quoted. 
I have deliberately refrained from collecting the numerous scattered records of odd 
crosses which exist in the various journals devoted to horticulture, because many 
of these are published either anonymously or by persons whose skill in performing 
the difficult operation of crossing in such a way as to be sure of the result may 
possibly be doubted. The evidence brought together rests upon the statements of 
men Avhose knowledge and skill are beyond question, and the only conclusion 
which can, I think, be fairly drawn from it is, that dominance of any of the 
characters mentioned is not an invariable attribute of the character, but that a 
cross between pairs of parents, such that the different members of each pair differ 
to the same extent in cotyledon colour or in similar characters, may in different 
