2. CARMICHÆLIA, R. Br.
Erect or depressed shrubs, some species attaining a height of 6–10 ft., others reduced to broad matted patches hardly rising more than an inch or two above the ground. Branchlets flattened or terete, grooved or striate, green. Leaves often absent, except in seedlings; when present deciduous after the flowers have fallen, 1-foliolate or pinnately 3–5-foliolate. Flowers small, in lateral racemes springing from notches on the edges of the branchlets, rarely solitary. Calyx campanulate or cup-shaped, 5-toothed. Standard orbicular, usually reflexed, contracted into a short claw. Wings more or less falcate, oblong, obtuse, auricled towards the base. Keel oblong, incurved, obtuse, shorter or longer than the standard. Upper stamen free, the others connate into a sheath. Ovary narrowed into a slender beardless style; stigma minute, terminal; ovules numerous. Pod small, coriaceous, narrow-oblong to almost orbicular, straight or oblique, compressed or turgid, narrowed into a short or long subulate beak; valves with the edges thickened and consolidated, forming a kind of framework called the replum, from which the faces of the valves come away; or in a few species the valves remain attached to the replum and the pod is indehiscent. Seeds 1–12, reniform or oblong; radicle usually with a double fold.
A very remarkable genus, confined to New Zealand, with the exception of one species found in Lord Howe Island. Its habit is peculiar, most of the species being leafless or nearly so when mature, the green flattened or terete branchlets (cladodes) performing the functions of true leaves. The structure of the pod is most exceptional, the margins of the valves and placentas being thickened and consolidated into a framework (replum), to which the seeds are attached. In dehiscence the faces of the valves either come away altogether from the replum, which may persist for a long time with the seeds hanging from it, or the valves may separate at one side or end, remaining attached at the other. In the four species constituting the section Huttonella the valves do not usually separate from the replum, which is frequently incomplete, and the pod is thus indehiscent. Had this character been constant, Huttonella might well have been kept as a distinct genus, as proposed by Kirk. But fruiting specimens of C. juncea in Mr. Colenso's herbarium show that the valves occasionally separate from the replum in that species, and Mr. Petrie informs me that the same thing occurs in his C. compacta.
The discrimination of the species is probably more difficult in Carmichælia than in any other genus in the New Zealand flora, and the student will find it almost impossible to name his specimens with accuracy until he has collected most of the species and become familiar with their characters. In most cases characters based upon the vegetative organs are by themselves useless. The leaves, when they can be examined, are singularly uniform; and the branchlets are not only highly variable in width, but may be flattened in spring and nearly terete in autumn. The flowers vary in size and colour in the different species, but present no important structural modifications. The pods afford the most trustworthy characters, and in several cases are alone quite sufficient for the identification of the species. The following analysis of the species is in many respects imperfect, and will doubtless require considerable modification. A really comprehensive and accurate account cannot be drawn up until the species have been carefully studied in the field at different seasons of the year, and in all stages of growth. It is specially important, in order to form a safe basis for future work, that flowering and fruiting specimens should be taken from the same plant.
