336 JOURNAL OF THE ROYAL* HORTICULTURAL SOCIETY. 
coloured varieties, two-thirds of the produce were ‘ Boule de Feu,’ and 
the remainder of very little value. 
I have come to feel such a certainty in my hybridisations that I can 
work like a painter in water-colours, by placing one colour over another, 
and not by mixing the tints; that is to say, to take a Gloxinia with deep 
violet-blue flowers, perfectly fixed as regards colour, and fertilise it by an 
equally well-fixed red one, and then not have any mixture of colours at 
all, but a superposition, a glazing of red over violet of very fine effect. 
In order to operate in this manner, one must be very sure about his 
plants. 
I think the preceding notes will suffice to thoroughly show the 
advantage to be gained by acting in fertilisation with plants of which 
the origin has been properly studied and the sources of parentage have 
been carefully preserved. 
Such methods are absolutely indispensable to all who are desirous 
of making fertilisations, and secure results which are not entirely due to 
chance. 
My friends, MM. Vallerand, have been able to carry to the utmost 
perfection the race of dotted and marked Gloxinias and the sub-varieties ; 
[ have brought, by dint of work and attention, the race of Crassifolia to 
a perfection which, I am entitled to think, has not been excelled. 
What I have written may still help those who wish to attempt to 
improve it further! 
