358 
ON CROSSES AND 
its pollen having reached the conservatory, though it cer- 
tainly was possible that it might have done so. Twelve 
seedlings were raised from the fruit, which was small, shri- 
velled, and quite deficient in juice, and those which have 
flowered not only approximated in colour to ccerulea, having 
no tinge of the red of its female progenitor, but were inferior 
to ccerulea in the beauty of the flower, and tenderer than 
that plant ; neither of which circumstances were likely to 
have occurred, if they had been derived from a second cross 
with ccerulea. In the same manner I have found that the 
seedlings from the crosses, between the scarlet G. cardinalis 
and the white or purplish G. blandus, are always disposed 
to degenerate from the colour of the more brilliant parent 
and approximate themselves to G. blandus, whether the 
scarlet cardinalis was the male or the female ancestor. It 
appears probable that this seeming disposition in fertile 
crosses to produce seedlings approaching to the least splendid 
of their parents, may arise from the effects of our climate 
upon them, which is more congenial to the duller coloured 
than to the brighter species ; in which case it would follow, 
that if the crosses were planted in the native soil and atmo- 
sphere of their more splendid parent, the same degeneration 
of colour would not take place. This is, however, a conjec- 
ture which I have no opportunity of verifying. I was led in 
some measure to form it, by having once observed the flowers 
of the hardy Nymphaea alba of a pale rose-colour, after a 
fortnight of unusual and intense heat in July, which ap- 
peared to point out why the genus Nymphaea, which is white 
in our latitudes, is found blue nearer the tropics, and red 
under their influence. This suggestion does not, however, 
account satisfactorily for the mule offspring, being inferior to 
the mules themselves generated in a similar situation ; but I 
have observed the seedlings from Hippeastrum Johnsoni or 
Regio-vittatum by its own pollen to have often a corolla both 
smaller and less brilliant than the mule plant itself, and this 
deterioration of the descendants may perhaps be in part 
attributable to the fertility of the mule being less vigorous 
and perfect than that of the original parents, when there 
exists some constitutional difference between them, which is 
the case in these three instances, Passiflora ccerulea being 
hardier than racemosa, Hippeastrum vittatum than regium, 
and Gladiolus cardinalis much more thirsty than blandus. 
I have already spoken of hybrid cactaceous plants of the 
