THE ORCHID WORLD. 
Cattleyas, and a very pleasing fact to be men- 
tioned is the special interest Mrs. Gratrix takes 
in looking after the wants of these delicate and 
youthful Orchids. Some really good crosses 
have been made and in due course first-class 
hybrids will probably flower. 
Mr. T. Brown, who has charge also of the 
17 acres of grounds, shows fully his capabili- 
ties as an expert Orchid grower, and the 
many rare and beautiful plants with which he 
is entrusted should act as a great incentive to 
his ever willing desire to still further improve 
their good qualities. 
G W. 
SOME IDEAS ON ORCHID HYBRIDISATION. 
By J. M. BLACK. 
{Continued from page 1 1 5.) 
ONE of the laws of inheritance is very 
well-known, and has been for untold 
ages, and that is that " breed will tell." 
This colloquial phrase sums up very neatly 
the secret which tlie Orchid hybridist who 
intends to get to the top must ever have in his 
mind. All successful breeders of animals and 
plants, whether versed in Mendelism or not, 
are aware of this cardinal fact, and are 
constantly applying it. Nature concerns 
herself with the preservation and perpetuation 
of the species, and does not aim at anything 
more ; her work is entirely socialistic, uncon- 
sciously levelling up and levelling down she 
stedfastly pursues the mean average. But 
this known law of inheritance consciously 
applied has given us better than the averag-e 
of unassisted nature, and its continued 
application will give us many more super- 
orchids. A moment's cogitation will satisfy 
any one with but even a very short acquaint- 
ance with Orchids that it will be futile and a 
waste of time and effort to produce from seed 
a type of Orchid that for all practical purposes 
is no better than a type that can be imported 
at little cost. It will be an expensive and 
long-drawn-out way of achieving something 
which when achieved will not be worth the 
trouble. There has been in the past a too 
eager desire to get seedlings at any hazard, 
and the strictly rigid selection of parent 
plants that I would counsel has not been too 
closely followed — if at all. The selection of 
parent plants should not be bounded by the 
frontier of the collection, and a fine variety 
should never be mated with an inferior one. 
Some growers of the old school were disposed 
to deride the hybrid, and were wont to 
describe it as " two good things spoilt," and 
that description was, and is still, too often 
justly deserved ; but it would oftener be 
nearer the mark to say " one good thing 
spoilt," the result of crossing a high-class 
variety of one species with a poor variety of 
another. But it is quite possible to have two 
good things spoilt in the hybrid through badly 
thought-out crossing, and most offences of 
this kind have occurred through crossing light 
and dark varieties with one another. As an 
illustration, let me take two familiar varieties 
of Dendrobium nobile, namely, nobilius and 
virginale, the former remarkable for its rich 
colour, and the latter for being pure Avhite, 
both in their way high-class, although now 
inexpensive Orchids. If one had a desire to 
raise a fine strain of D. nobile the mating of 
virginale with nobilius, although both among 
the very best of nobile varieties, could only 
result in an inferior variety, as the colours 
would neutralise each other — nature finding 
the mean average — ^and we would thus 
succeed in spoiling two good things. Dendro- 
bium nobile can be imported in big clumps 
at a cost low enough to be sold at a very 
nominal figure, so that the raising of it from 
seed to attain only the mean average would 
be a waste of time. The result of the ferti- 
lising of D. n. nobilus with D. n. virginale 
would, strictly speaking, not be a hybrid, 
although there is much more dissimilitude 
between these two varieties of one species 
than between many species that are botani- 
cally recognised as distinct, and they will 
serve very well to further illustrate my 
meaning. Our object is, say, to raise a batch 
of Dendrobium nobile better than it would be 
