STUDIES ON THE GENETICS OF FLOWER-COLOURS, ETC. 
97 
Flower-colours on one and the same individual are often more or less 
different in early and late summer, and also at different hours of one day. 
Colours in the above Table are those recorded in July and generally soon 
after the opening of flowers. 
Factors. 
Before proceeding further I will mention the factors concerned in the 
colour production of varieties used by me. Of course the action of all these 
factors has been discovered only after many breeding experiments have been 
carried on, but in my description I will take the opposite course, because I 
will mention the action of the factors at first, and then go. on to the crossing 
experiments. 
The factors producing the flower-colours of Portvlaca are as follows :— 
1. C, fundamental factor for the production of any colour, copiant being 
white ; C alone, either in one or double dose, makes flower orange ; 
2. G ( gilvus ), changes the orange colour produced by C into yelloio ; 
3. It changes the orange colour produced by C into red ; 
4. B which I will call blueing factor changes the red colour produced 
by the co-operation of C and It into magenta. B, without B, even in presence 
of C, has no blueing action. 
The factors specially concerned in the genotypic constitution of flesh- 
coloured and white-III races will not be considered in the present paper. 
All my breeding experiments given below lead us to the conclusion that 
the varieties used by me may be expressed in respect to the colours of petals 
by the following genetical formulae, if we adopt the presence-and-absence 
hypothesis :— 
1. Orange CCggrrbb (PI II, fig. 3). 
2. Yellow CCGGrrbb ( „ „ „ 4). 
3. Red CCggBBbb ( „ „ „ 2). 
4. Magenta CCggBBBB ( ,, „ ,, 1). 
5. White-I ccggrrbb ( „ „ „ 8;. 
6. White-II ccggBBBB ( „ „ „ 7). 
In describing the cross experiments below the letters gg are omitted in 
genetical formulae, except in the Cross II. 
