21 



each packet containing a large number of these grains. The anther 

 consists of two widely separated cells, and each cell contains a 

 pollen- mass, or, as it is termed, a pollinium. 



The Arethuscc have a good representative in Britain in Ceplialan- 

 thera grandiflora. This orchid is self-fertilised and does not possess 

 a rostellum. The pollen is unlike most orchids and is not in masses, 

 but is friable, with the grains held together by small elastic threads. 

 The grains adhere to any object readily, and when the anther opens 

 whilst the flower is in bud it expels the pollen to fall on the stigma. 

 Darwin considered this a degraded orchid, according to its fertilising 

 arrangement. 



The / r andecs, which contains a large proportion of both old- and new- 

 world orchids of our glasshouses, and the Epideudreix, which includes 

 the most beautiful new-world genera, have such a diversity of 

 fertilising structure that it would fill many volumes to describe them 

 in detail, but all are adapted for "crossing " by insects. 



The Cypripediece are widely separated from all other orchids, show- 

 ing that there must have been many connecting links which have in 

 ages past become extinct. The great difference in the fertilising 

 structures of the Cypripediece is that the single anther, which is 

 present in all other orchids, is here rudimentary (to use Darwin's 

 own words), and is represented by a singular shield-like projecting 

 body, deeply notched and hollowed out in its lower margin. There 

 are two fertile anthers which belong to an inner whorl represented 

 in ordinary orchids by various rudiments. 



In the Vaudeic there is only one large disc in the stigma, where 

 the pollen is received, but the genus Atigrcecitm is an exception and 

 has two. In the Ophrydecv exactly the reverse is the case. The 

 usual structure is two separate glands to the stigma, except in the 

 genus Orchis, which has only one gland. 



As the orchid family, speaking generally, is so perfectly adapted 

 for cross-fertilisation, experiments were made about forty years ago 

 in hybridising, not only different species in a genus, but by crossing 

 species in different though allied genera. The result was at once 

 successful, though the rearing of the seedlings up to flowering plants 

 was, and is now, a tedious process. The early hybrids were with 

 Limatodes (Calanthe) rosea (seed parent) crossed with Calanthe vestita 

 (pollen parent), giving Calanthe veitchii, Ltzlia crispa (seed parent) 

 crossed with Cattleya mossice. giving Lcelio-cattleya exoniensis and 

 several others. Crosses between the genera Lcelia and Cattleya, and 

 between different species within these two genera, have always been, 

 from the first, sought after by orchid fanciers, the brilliant colourings, 

 large size and fine texture of the flowers sufficing to make up a 

 beautiful object for anyone's attention. But with many other genera 

 that have been freely hybridised there has been great demand only for 

 a time, while it was, so to speak, the fashion. Hybrid Dendrobiums 

 were, five and twenty years ago, largely grown and found a ready 

 sale, often at very long prices, but to-day they find little favour with 



