184 
R. Ruggles Gates . 
var. perlutesccns, which is yellow throughout, that the yellow palate 
was dominant. Thus yellow palate is dominant in one species and 
recessive in the other—a quite unexpected result. It is evident 
that dominance or recessiveness cannot safely be used in deter¬ 
mining which is the original or parental type. Even though these 
are parallel mutations, the alternation in dominance probably 
depends on an effect of the other specific differences. 
In Digitalis purpurea several varieties, evidently mutations, 
have been studied. Saunders (1911), Shull (1912) and others 
have bred the form called by Chamisso heptandra in 1826. The 
flowers are characterized by dialysis and staminody of the corolla, 
giving a flower with a petaloid upper lip and seven stamens. But 
many flowers of a heptandrous plant may show the abnormality 
in varying degrees, like a wave of reversion advancing up the plant. 
This variety is a recessive in inheritance. 1 A form nudicaulis with 
smooth stems and leaves less hairy on the upper surface is found by 
Miss Saunders (1918) to be a simple dominant to the type. Although 
this condition corresponds to the half-hoary type in stocks, yet in 
the latter it is due to several factors and behaves as a Mendelian 
recessive to hoary. This is another example of externally similar 
characters which are wholly different in their genetic relationships, 
and it shows that certainty regarding variety relationships can 
only be attained by actual breeding experiments. D. purpurea 
nudicaulis occurs wild in various parts of England mingled with 
pubescens t the hairy type. The latter is usually more abundant and 
the indications are that nudicaulis has originated from it as a 
mutation. That a new dominant character can arise by a mutation 
is known from the case of CEnothera rubricalyx among plants or 
such mutations as bar eye in Drosophila among animals. 
From an almost unlimited number of cases of new varieties or 
forms probably originating as mutations in wild species and often 
already in process of spreading, we may select a few almost at 
random. 
1 An exactly parallel mutation hasdong been known in a North American 
member of the Ericaceae, Kalmia latifolia. In 1871 Asa Gray (Am. Nat. 4 : 373) 
described a plant from South Deerfield, Mass, showing dialysis with staminody. 
It was grown in the Arnold Arboretum and figured by C. S. Sargent (Garden 
and Forest 2 : 452). It seeds freely, but apparently its inheritance has not 
been tested. In 1909 Stone (Rhodora 11 ; 199) independently reported the 
same form (var. polypetala Nicholson) from the roadside in Leverett near Mt. 
Toby, and Rehder (Rhodora 12 : 1, 1910) adds a number of varieties, evidently 
mutational, of this species. Fernald (Rhodora 15 : 151, 1913) records near 
St. John’s, Nfld„ a considerable colony of K. angustifolia with white corollas 
(forma Candida), This form was also found at Sherborn, Mass., and evidently 
occurs as a sporadic negative mutation which will spread if left to itself. 
