378 Transactions. — J lot any. 



Art. XXXVI. — Notes on Carmichaslia, with Descriptions of new Species. 

 By T. Kirk, F.L.S. 

 [Read before the Wellington Philosophical Society, 13th February. 1881.] 

 Plates XXX.-XXXIIL 

 The genus Carmichatlia comprises certain species which are tolerably con- 

 stant in then- vegetative characters, and can be identified in all states with 

 little difficulty, but on the other hand it contains species of which the indi- 

 viduals exhibit an amount of variation scarcely surpassed by any genus in 

 our variable Flora. The habit of the plant may be lax or compact ; the 

 branches and branchlets may be glabrous or more or less pubescent ; they 

 may be terete, plano-convex or much compressed, and when compressed 

 may vary from one-twentieth to one-fourth of an inch in breadth ; they 

 may be leafless or foliaceous, while unifoliolate and pinnate leaves may be 

 found on the same branch. The inflorescence may consist of two or three 

 few-flowered fascicles, or of fascicles or short racemes crowded into false 

 whorls, or of many-flowered lax racemes, and the pedicels may be glabrous, 

 pubescent, or pilose. 



The most constant characters are afforded by the fruit which varies 

 widely in the different species. In the typical form of the pod dehiscence is 

 effected by the separation of the valves from the placenta, leaving the latter 

 as a framework (replum) carrying the seeds, but in some species the replum 

 itself splits down the middle, one-half being attached to each valve ; in 

 others one valve becomes partially or completely separated from the replum 

 which remains attached to the opposite valve. The replum varies con- 

 siderably in the different species, being least developed in C. crassicaulis. 

 Dehiscence usually commences from the base, but not invariably. 



The form of the pod exhibits marked characters in the different 

 species, and nearly always affords the safest characters for identification, 

 however much the species may vary in its vegetative characteristics. 

 With a few exceptions characters derived from the flowers are less satis- 

 factory. 



I do not propose to consider the causes of variation in Carmichalia 

 beyond recording the fact that in 0. odorata, C. pilosa, and G, flagellifor- 

 mis — the species in which it is most excessive — it is often caused by soil, 

 situation, age of the plants, etc. 



In the following paper I purpose describing two new species, and sup- 

 plying some omissions in the characters of those with which we are already 

 acquainted, from descriptions in the " Handbook of the New Zealand 

 Flora " and elsewhere. 



