148 Half-hoariness in Matthiola 



here'. It will suffice to emphasise the fact that mere inspection of the 

 individual in flower will not enable us to distinguish between the 

 different forms included in any one of the three groups. There is 

 nothing visibly different in the different classes of whites and creams ; 

 and among sap-coloured plants, individuals of the same colour and 

 surface character may belong to any class within the group, or, on the 

 other hand, plants of different colours (purple, red, flesh, etc.) may all 

 belong to one class. The only method of identification at present 

 available is the slow and laborious one of cross breeding, which 

 necessitates the separate testing of each indi\ddual. The commercial 

 material from which, from time to time, a fresh start had to be made 

 was found to give very uniform results. Matings between hoary and 

 glabrous types constantly gave all hoary= in jP, and a mixed offspring in 

 F2 in the ratio of 3i/ : IG. When two glabrous forms were intercrossed 

 a hoary F^ was sometimes obtained in cases where both parents were 

 non-sap-coloured, and again where one was sap-coloured and one not. 



1 "Further Contribution to the Study of Hoariuess in Stocks (MattUiolii)." Proc. 

 Roy. Soc. B. Vol. lxxxv. 1912. 



^ During the progress of these experiments I have become convinced that when the 

 hoary parent is homozygous this result invariably obtains. In the first set of experiments 

 set out in Report 1 (loc. cit.) a certain number of cases are given in which glabrous 

 individuals were recorded among the I'\ plants from the mating glabrous ? x hoary J which, 

 if genuine, suggested the possible occurrence of parthenogenesis. (See Report I, Table II. 

 p. 38 and Table XIV. p. 83 in which the reference numbers of the particular experiments 

 are quoted under "aberrant cases.") That these cases are not genuine exceptions .md are 

 not to be explained as due to parthenogenesis I now feel no doubt. I believe them to be 

 due to the omission of some precaution which later experience led me to adopt, though 

 I am still at a loss to suggest the particular cause of the experimental error. 



The same remark applies to the case of Salvia Horminum. In the account of the 

 investigations of the relations of the pink and white varieties to each other and to the 

 violet type, which appears in Report II, two exceptions are recorded to the dominance of 

 pink to white, and it is there suggested that these two exceptions may be due to 

 experimental error. This we may safely hold to have been the case, for the factor relations 

 in the case of Salvia are of the simplest type. The pink variety owes its colour to the 

 presence of a factor P. The violet type contains the colour factor P and in addition a 

 factor B which turns the pink colour blue. Whites lack the colour factor P and hence B 

 which was present in the plants employed is ineffective. This view of the presence of 

 a single colour-producing factor upon which another factor changing the class of colour 

 may be superposed, which therefore regards the white as lacking this same colour factor 

 rather than as containing a distinct factor IC, leads us to expect the now familiar ratio 

 gV -.SP: 4IC in /-'^ from P\ IV (i.e. bPxBp) and not, as originally suggested, 2V : IP AW. 

 The numbers actually recorded are, as will be seen on reference to the Tables, in closer 

 agreement with the correct interpretation. The "presence and absence" view is now 

 so familiar that this note would hardly have been called for were it not that no actual 

 reference to this particular case putting it in line with many others happens to have been 

 made in the course of the later work. 



