trials with fleshy rhizomes; or epiphytes with creeping stems and rhizomes. Leaves simple, alternate, often distichous, sheathing at the base. Flowers solitary or in loose or spicate racemes, hermaphrodite, zygomprphic. Perïanth epigynous, composed of 6 segments in 2 whorls, which are sometimes similar, or more often with the lateral petals paired and the lateral sepals paired and the middle segments of both whorls modified to a greater or less extent, the middle petal (labellum) particularly, often becoming extremely complicated in structure. The ovary is usually twisted through 180°, bringing the labellum into the abaxial position. Theoretically there are 6 stamens in 2 whorls; the 2 lateral of the outer whorl confluent with the middle petal to form the labellum, the third of the outer whorl fertile at the apex of the column; the 2 lateral of the inner whorl modified into various appendages (column-wings); the third of the inner whorl confluent with the front of the column. (In Petalochilus it appears as a sigmoid appendage at the base of the column.) Anther 2-celled, introrse, opening by a slit lengthwise; pollen granular, or generally agglutinated into mealy or waxy masses (pollinia), which may be extended at one end into a sterile caudicle. Pollinia free in the anther-cells or loosely united. Ovary inferior, 1-celled with 3 parietal placentas. Stamens and style combined to form a variable structure (column) in the centre of the flower. Stigmata 3, the lateral 2 fertile, the third modified to form the rostellum, to which the pollinia are frequently attached. Ovules very numerous, minute, anatropous. Fruit a capsule, opening by longitudinal slits. Seeds very numerous, minute, without endosperm. Embryo not differentiated.

It must be made clear that the above description is only intended to cover New Zealand genera. In other members of the family the leaves may be opposite, or the flowers polygamous or monoecious. There may be 2 fertile stamens or 3 fertile stigmata, or the ovary may be 3-celled with axile placentas, or the seed may be winged.

(C) The following system of classification is based upon Schlechter (Notizbl. Bot. Gart. Berl., 9, 1926, 567) and was prepared for this paper by Mr. Leslie A. Garay, of the University of Toronto, Canada. It is a considerable advance upon the system at present used in Australia (refer Rupp, Orch. N.S.W., 1943, 2) since it gives a more natural position to Calochilus, Microtis and Prasophyllum.

(D) In Trans. R.S.N.Z., 76, 1946, 58, the writer endeavoured to show how close was the relationship between the orchid floras of Australia and New Zealand. This may be accounted for when we consider the various trends of distribution which have and are still affecting the South Pacific floras.

(a) It was originally considered (Rupp and Hatch, Proc. Linn. Soc. N.S.W., 70, 1945, 57) that 7 genera showed signs of having differentiated in a south-to-north direction. This opinion, for which the present writer was alone responsible, has had to be revised from time to time until at present the only genera showing definite signs of south-to-north evolution are Pterostylis (Hatch, Trans. R.S.N.Z., 77, 1949, 234) and Thelymitra. These genera are presumed to have originated in the paleozelanic continent which the evidence leads us to believe extended considerably to the southward of the present position of New Zealand (but hardly to Antarctica) and also to Tasmania, during early Cretaceous times.

(b) The second trend was presumably an overland movement by way of Malaya, Java, New Guinea, New Caledonia, Lord Howe Island, etc. This brought us the truly Asiatic genera Corybas, Gastrodia, Spiranthes, Microtis, Dendrobium, Bulbophyllum, Sarcochilus. (Earina, originating in New Caledonia, probably came to New Zealand by this route also.)

(c) There is a distinctly powerful west-to-east, probably windborne movement which brings East African orchids to Western Australia, Eastern Australian orchids to New Zealand and New Caledonia, and thence to the Chathams and Polynesia. This gave us the Australia-originating genera Caladenia, Lyperanthus, Chiloglottis, Townsonia, Adenochilus, Acianthus, Orthoceras, Prasophyllum, Caleana and Galochilus, and in addition odd Australian species of genera which normally arrived by other routes (Pterostylis, Thelymitra, Corybas, Gastrodia, Microtis). That these have crossed the Tasman in dust storms in the past cannot be denied, but we must also consider the possibility of Australian orchids continuing to do so at the present time. How else to account for the isolated occurrence of Townsonia on Mt. Ruapehu, Pterostylis nutans on the Whangaparoa Peninsula, or Pt. mutica on the Waiouru Hills? And lastly the occurrence of Thelymitra venosa cyanea, which occurs in New Zealand only in Stewart and the Auckland Islands, at Kaikohe in the far north?

(d) And lastly the tendency to local evolution which produced Petalochilus from Caladenia, and Aporostylis from a combination of Caladenia and Chiloglottis, and which led and is still leading to the differentiation of endemic species in the larger genera.

These trends apply equally of course to the flora as a whole, particlarly (a), (b) and (d), but section (c) applies more directly to the orchids than to other families because the minute seeds are more easily carried in dust storms and by high winds. Also these seeds