would have suspected that these and many other small details of structure were of the highest importance to each species; and that consequently, if the species were exposed to new conditions of life, and the structure of the several parts varied ever so little, such small details of structure might be modified by natural selection. These cases afford a good lesson of caution with respect to the importance, in other organic beings, of apparently trifling particulars of structure. It may naturally be inquired, why do Orchids exhibit so many perfect contrivances? From the observations of C. K. Sprengel, and from my own, I am sure that many other plants offer, in the means of fertilisation, analogous adaptations of high perfection; but it seems that they are really more numerous with Orchids than with most other plants. To a certain extent the inquiry can be answered. As each ovule requires at least one, probably more than one, pollen-grain,[1] and as the seeds produced by Orchids are so inordinately numerous, we can see that large masses of pollen would have to be left on the stigma of each flower. Even in the Neottee, which have granular pollen, with the grains tied together only by weak threads, I have observed that considerable masses of pollen are generally left on the stigmas. Hence we can perhaps understand the use of the grains cohering in large waxy masses, as they do in so many tribes, so as to prevent waste in the act of transportal. Most plants produce pollen enough to fertilise several flowers, even when each flower has several stigmas. But as the two confluent stigmas of Orchids require so much pollen, its elaboration, if proportional in amount to that in most other plants, would have been extravagant in the highest degree, and exhausting to the individual. To save this waste and exhaustion, special and admirable contrivances were necessary for safely placing the pollen-masses on the stigma; and thus we can partially understand why Orchids have been more highly endowed in this respect than most other plants.
The simple fact that many Vandee have only two pollen-masses, and that, from their coherence, some Malaxee have only a single pollen-mass, necessitates that extraordinary pains should have been taken in their fertilisation, otherwise these plants would have been barren. The existence of a single pollen-mass, so that all the pollen-grains from one flower cannot possibly fertilise more than a single stigma, occurs, I believe, in no other plants. The case is partially analogous with that of seeds: many flowers produce a multitude of seeds, several profuce a single seed; very many flowers produce a countless number of pollen-grains; some Orchid-flowers produce, as far as the power of fertilising other flowers is concerned, a single pollen-mass, though really consisting of a multitude of pollen-grains.
Notwithstanding that such care has been taken that the pollen of Orchids should not be wasted, we see that throughout the vast Orchidean order,—including, according to Lindley,[2] 433 genera, and probably about 6000 species, the act of fertilisation is almost invariably left to insects. This assertion can hardly be considered rash, after the examination of so many British and exotic genera scattered through the main Tribes, which generally have a nearly uniform structure. In all plants in which insects play an important part in the act of fertilisation, there will be a good chance of pollen being carried from one flower to another. But in Orchids we have seen numerous adaptations, such as the movements of the pollinia after their removal in order the acquire a proper position,—the slow movement of the labellum or rostellum to allow the entrance of the pollen-masses, the separation of the sexes in some instances—which render it certain that in these cases the pollen of one flower or of one plant is habitually transported to another flower or plant. As this transportal increases the risk of loss, it necessitates and still further explains the extraordionary care bestowed on the contrivances of fertilisation.
Self-fertilisation is a rare event with Orchids. In Cephalanthera grandiflora it occurs, but in a very simple degree; and the early penetration of the stigma by the flower's own pollen-tubes seems to be fully as much determined by the support thus given to the pillars of pollen, as by the production of a small proportion of seed: certainly the fertilisation of this Orchid is aided by insects. In some species of Dendrobium self-fertilisation apparently occurs, but only if insects accidentally fail in removing the flower's own single pollen-mass. In Cypripedium, the Frog Orchis, and perhaps in a few other cases, it will depend on the manner (at present unknown) in which insects first insert their probosces by the one or the other entrance, whether the flower's own pollen, or that of another flower, is habitually placed on the stigma; but in these cases there will assuredly always be a good chance of the stigma being fertilised by pollen brought from another flower. In the Bee Ophrys alone, as far as I have seen, there are special and perfectly efficient contrivances for self-fertilisation, combined, however, in the most paradoxical manner, with manifest adaptations for the occasional transport by insects of the pollinia from one flower to another, as in the other species of the same genus.
