Parallel Mutations. 47 



For example, each time lata appears as a mutation from lamarckiana 

 the change involved must be the same. Again, in the sense in 

 which the term originated, (E. biennis lata, lamarckiana lata and 

 suaveolens lata are all parallel forms. Going farther afield the 

 cruciata forms from lamarckiana and biennis, if independent in 

 origin, are obviously parallel, but a number of wild species (shortly 

 to be considered) which are unrelated to each other have cruciate 

 flowers as a specific character. Some of these may have a common 

 descent from a cruciate ancestor, others have almost certainly 

 originated independently as parallel mutations. Nor is it certain that 

 the change involved is always the same. The Drosophila work has 

 shown that two mutations may be so closely alike as to be scarcely 

 distinguishable, and yet have different relations in the germ-plasm. 

 The conception of parallelism is perhaps even more fertile 

 when applied to the comparison of variations in different genera. 

 A cruciate mutation has appeared in Epilobium and a similar 

 variation probably occurs in other genera. Again, the phenomenon 

 of doubling in flowers, to be considered later, shows that doubling 

 may occur as parallel mutations in wild as well as cultivated species 

 of many families. The same of course is true of white varieties, 

 which almost any species of flowering plant may produce. 

 Laciniation of leaves, maximum anthocyanin development, nanism, 

 lack of chlorophyll (though such a condition may arise from various 

 genetic causes), doubling, and a number of other characters may be 

 looked upon as probably at any rate parallel mutations. Thus (E. 

 rubricalyx and the copper beech may be parallel in that they 

 involve in their origin the same kind of germinal change. But this 

 cannot be assumed with safety until it has been proved by breeding 

 experiments with each. Thus some forms of flower-doubling are 

 dominant and some recessive, and white varieties may originate in 

 a number of ways. The same is true of gigantism. The gigas 

 mutations from four species of CEnothera (Table II) show exactly 

 the same peculiarities of stoutness, increase of cell size, 4-angled 

 pollen grains and tetraploid chromosomes, so that the change is 

 obviously the same in all cases. But in other genera gigantism may 

 occur with or without tetraploidy. This whole subject will be 

 considered further in connection with the variations of wild species 

 of plants and animals. 



