IN THE TROPICS. 
345 
VIII.— DE VRIES' ANALYSIS OF FLORAL 
COLOURS . 
In addition to his experiments with peas and with 
Hieracium, Mendel made crosses also between certain 
strains of Phaseolus, in order to test whether his conclu¬ 
sions held good with those plants as well. In the case of a 
certain cross between Phaseolus vulgaris and P. nanus he 
found the result to be in perfect agreement with those 
previously described for Pisum crosses. 
Mendel next crossed Phaseolus nanus, which had white 
flowers and seeds, with pollen from P. multiflorus, the flowers 
of which were purple-red, whilst the seeds had purple 
spots on a peach-blood-red ground. The hybrid plants 
(Fjl) resembled the pollen parent closely. In F 2 only 31 
plants reached the flowering stage. These showed in certain 
other characters almost exactly the ratio 3 : 1 between 
dominant and recessive allelomorphs. Only 1 plant,however, 
showed white flowers and seeds, although eight were to be 
expected if the case was one of crossing between simple 
allelomorphic characters. The plants with coloured flowers 
showed a wide range of character in this respect, being of 
various grades of purple-red to pale violet, and the colour 
of the seed coat was no less varied. In the further genera¬ 
tions only a small number of plants could be raised. 
Mendel suggested that the small proportion of plants with 
white flowers could be explained by supposing the dominant 
colour of the flowers of Phaseolus multiflorus to be made up 
of two or more entirely independent colours, A 1 and A 2 . 
Then on crossing by a, we get (A x a + A 2 a) ; more properly 
written A x A 2 a, as Bateson points out (2). From the gametes 
AjA 2 , Aja, A 2 a, and a, there arise in F 2 the combinations : 
(A, -f 2 A* x a + a) (A 2 + 2A 2 a + a), or : 
1 Aj A 2 2 A,a A 2 1 A 2 a 
2 A, A 2 a 4 A,a A 2 a 2 A 2 aa 
1 A,a 2 A t aa 1 aa 
and Mendel suggested that each of these combinations denotes 
