366 Menders Experiments 



The hybrids had the greatest similarity to the pollen 

 parent, but the flowers appeared less intensely coloured. 

 Their fertility was very limited ; from seventeen plants, 

 which together developed many hundreds of flowers, only 

 forty-nine seeds in all were obtained. These were of 

 medium size, and were flecked and splashed similarly to 

 those of Ph. multiflorus, while the. ground colour was not 

 materially different. The next year forty-four plants were 

 raised from these seeds, of which only thirty-one reached 

 the flowering stage. The characters of Ph. nanus^ which 

 had been altogether latent in the hybrids, reappeared in 

 various combinations ; their ratio, however, with relation 

 to the dominant plants was necessarily very fluctuating 

 owing to the small number of trial plants. With certain 

 characters, as in those of the axis and the form of pod, it 

 was, however, as in the case of Pisum, almost exactly 1:3. 



Insignificant as the results of this experiment may be as 

 regards the determination of the relative numbers in which 

 the various forms appeared, it presents, on the other hand, 

 the phenomenon of a remarkable change of colour in the 

 flowers and seed of the hybrids. In Pisum it is known 

 that the characters of the flower- and seed-colour present 

 themselves unchanged in the first and further generations, 

 and that the offspring of the hybrids display exclusively the 

 one or the other of the characters of the original stocks. 

 It is otherwise in the experiment we are considering. The 

 white flowers and the seed-colour of Ph. nanus appeared, it 

 is true, at once in the first generation \_from the hybrids] 

 in one fairly fertile example, but the remaining thirty 

 plants developed flower-colours which were of various 

 grades of purple-red to pale violet. The colouring of the 

 seed-coat was no less varied than that of the flowers. No 

 plant could rank as fully fertile ; many produced no fruit 

 at all ; others only yielded fruits from the flowers last pro- 

 duced, which did not ripen. From fifteen plants only were 

 well-developed seeds obtained. The greatest disposition 

 to infertility was seen in the forms with preponderantly 

 red flowers, since out of sixteen of these only four yielded 

 ripe seed. Three of these had a similar seed pattern to 

 Ph. multiflorus, but with a more or less pale ground colour; 

 the fourth plant yielded only one seed of plain brown tint. 



