EXPERIMENTS IN THE HEREDITY OF PEAS. 



487 



-would generally be sufficient to secure both the dominant and recessive 

 types. With regard to the precise nature of the dominant and recessive 

 characters in the second generation, no appreciable difference can be 

 perceived in the respective seeds of the different experiments, so that we 

 will take the seeds of Experiment 5 to represent them all. 



The wrinkled recessives in Experiment 5 are practically indistinguish- 

 able from the seeds of the pure recessive grand-parent ' British Queen,' as 

 far as shape is concerned ; all are angular, squared, flattened, with deep 

 dents and broad wrinkles. 



The Round dominants in Experiment 5 are all unmistakably round, 

 with no intermediate forms tending towards the wrinkled recessives : they 

 are, however, slightly variable in their roundness and smoothness. For 

 convenience they may be divided into three groups, though, as a matter 

 of fact, the groups are almost continuous. 



In order to enable others to identify and gain a clear idea of these 

 a ree groups, I have compared them as closely as possible with Professor 

 Weldon's photographs of seed shapes in Peas (1902, pi. II., figs. 1-18), 

 and in quoting the figures by number, where there are two peas in 

 Professor Weldon's groups, I will term them respectively A and B, reading 

 from left to right, in the same direction as the numbers. 



Out of 100 full-sized round seeds raised in Experiment 5, 9 may be 

 referred for shape alone to fig. 2A, and for smoothness alone to fig IB, 

 all being nearly round and perfectly smooth : 72 may be referred to fig. 

 2A for both shape and pitting, though in all, the pitting is less exten- 

 sive and pronounced : the remaining 19 may be referred to fig. 2B, 

 being slightly irregular in shape, with shallow dents. It will be noted 

 that 72 per cent, of the dominants of the second generation are similar to 

 the dominant hybrids of the first generation, while the rest comprise a 

 small percentage, varying on either side of the majority, and it seems 

 probable that, if 100 seeds of the first generation could have been 

 compared together (instead of the 19 actually compared), all these three 

 types would have been found in the same proportions as in the 

 second generation. In any case it seems likely that the slight im- 

 perfections of roundness and smoothness in both generations are due 

 to the action of similar causes. That one of these causes is not the 

 influence of the recessive character (causing incomplete dominance) is 

 suggested by the fact that none of the dominants in the second generation 

 are exactly like the pure dominants, and yet by Mendel's law one-third of 

 them should be pure dominants ; so that altogether it appears fairly clear 

 that in both the first and second generations the imperfections of round- 

 ness and smoothness are in some measure due to the conditions of life 

 rather than to heredity. It was further observed that the 9 per cent, 

 perfectly smooth seeds in the second generation were all smaller in size 

 than the others, and the 19 per cent, which were the most irregular were 

 the largest seeds of all ; evidently, therefore, size is also a factor to be con- 

 sidered, and it is hoped that an examination of the third generation will 

 determine this difficult and complicated question. 



With regard to the size of the seeds of the second generation, they are 

 sensibly smaller than those of the first generation. As a matter of fact, if 

 the size of the seeds is a purely maternal character, as it certainly appeared 



