118 JOUEXAL OF THE KOYAL HORTICULTUEAL SOCIETY. 
This last instance of adaptation to dispersal by animals which I 
have to mention well illustrates one of the leading conclusions drawn 
from such facts as those which I have put before you, viz. the probable 
specialising of such adaptations from more primitive and more indifferent 
adaptations. 
"We have seen that, though reducible to a comparatively small 
number of types, there are many very different mechanisms to secure 
seed-dispersal. Most types of mechanism are represented by members 
of widely separated orders : few^ of them are characteristic of whole sub- 
orders, hardly any of whole orders. Such genera as Geranium^ Viola, 
and Trifolium, and such orders as the Compositae, Crucifer^e, and 
Umbelliferte, exemplify the occurrence of a variety of mechanisms in 
allied plants. Considerable as is the variety of dispersal mechanisms as 
a whole, it is not as great as the variety of pollination mechanisms in 
the flower. These conclusions lead to the further inference that these 
dispersal mechanisms are on the whole of more recent acquisition in the 
vegetable kingdom, geologically speaking, than at least the more common 
of the floral adaptations. Kerner has argued convincingly that many 
adaptations are either still of an indifferent character or may at least 
easily be imagined to have had an indifferent origin. Small seeds, for 
example, as in Orchids, may at first have been the result of late develop- 
ment, or of mere economy of nutrition, and lend themselves as readily to 
simple censer action" as to dispersal by wind. Hooks and bristles, 
again, may be partly the result of purely nutritional causes, and may 
certainly serve, or have served, as protections against "unbidden guests," 
such as leaf-eating or honey-stealing insects, as well as, or before, serving 
to aid in dispersal by animals. Dispersal mechanisms afford, I think, 
some examples of progressive development in elaboration as instructive 
as Darwin's illustration of the possible origin of leaf-tendrils by the 
twining petioles of Clematis. Sir John Lubbock cites the case of the 
North American genus Thysanocarpus, belongiDg to the order Cruciferae, 
one living species of which, T. laciniatus, has a winged or margined 
siliqua ; whilst another, T. cnrvijjes, has considerably larger wings ; a 
third, T. radians, has them broader; and a fourth, T. elegans, has them 
so thin as to have become perforated. I have already alluded to the 
gradations presented by the wings of the seeds of different species of the 
genus Finns, and I would only furthur remark on the contrast afforded 
by the comparatively thick, heavy, and unformed wings of the samara of 
our common hedge Maple {Ace}' campesirc) and the beautifully curved 
and delicately thin vanes of other species, such as that familiar exotic 
the Sycamore [Acer Pseudo-platanus). Between these extremes there is 
a large series of intermediate forms. Considerations of time have 
caused me to intentionally say little on the present occasion as to the 
results of the seed- dispersal with which I have been dealing; but there 
is one interesting point to which, in conclusion, I wish to direct your 
attention. The epiphytic habit, whether it be exhibited by the Ferns, 
Gooseberry bushes, or various trees which may be found growing in the 
tops of our English pollard Willows, or by the multitudinous Orchids, 
Aroids, Tillandsias, &c., of the tropical forest, must be entirely dependent 
upon dispersal by wind or animal agency. 
