growing and are several years old before they begin to bear flowers. Even so, the effort seems to be too much, for it is quite a common thing for trees to rest a season and bear no flowers. Also, trees growing close to the shore usually blossom earlier than trees growing a few miles inland.

Leaves.—The young seedlings bear leaves of a full adult size, each leaf, however, with one notch at the apex (Fig. 1). This seems to be a slight approach to the heterophylly exhibited by so many of our New Zealand plants. The apex of the adult leaf is sub-acute,

almost rounded, as seen in Fig. 1, where a seedling is shown bearing two juvenile and two adult leaves. All leaves when fully grown are dark green, smooth, and glabrous, possessing on both surfaces numerous small depressions like pin-pricks. The young leaves in the bud are very interesting, and a description of their development follows.

The young stem shows at the apex typical stem structure. A very large number of vascular bundles is present. The young leaves

which cover the growing point have thick bases which ensheath about two-thirds of the circumference of the stem, and a leaf base in its lowest part where its cells are indistinguishable from the cells of the stem contains already the three fibro-vascular strands which are described by Boodle and Fritzsch as being characteristic of the genus. They are large strands and connect directly with the stem stele. As the blade of each leaf-rudiment becomes separated from the stem it always happens that one margin becomes separated earlier than the other, and if the leaves are considered in the order of their development along the genetic spiral it is always the margin that on a conventional genetic spiral would be the nearer to the apex that separates

first. The median bundle divides into five and the developing leafrudiments show the seven bundles which are characteristic of the genus arranged in horseshoe shape. The “blades” of the leaf-rudiments are not expanded like laminae generally, but are prismatic, with three sides (Fig. 2), and if these blades are examined they are found to possess on their upper surfaces curious hairs, but none on

A Cross Section of a Vegetative Bud.
At a one of the extreme fibro vascular bundles is dividing again.

their lower surfaces. As development goes on the prismatic shape becomes less evident, and at the same time there develop on the convex lower surface two indentations or clefts (Figs. 2 and 3). These clefts become carried towards each other parallel to the upper surface with the result that the blade splits into two halves, an inner and an outer. It so happens that the inner half has the two extreme fibro-vascular bundles, one near each end, and these two bundles are two of the three bundles which passed out originally from the stem and which never divided. This inner half is flattened like a lamina, but it has a narrow region along the middle, where there is no fibrovascular strand. Its widest portions are where the fibro-vascular strands lie (Fig. 3), and it possesses so far all the hairs. The outer half is prismatic, in cross section triangular, with two abaxial short sides, and the long adaxial side adjacent to the abaxial side of the inner half leaf-rudiment from which it was split off (Fig. 3).

After this stage in development the inner half develops no more hairs, both halves increase and grow in length at the same rate until they reach a length of a centimetre or more, and a young shoot now presents the appearance of having two leaves inserted together on the stem (at the same spot), one inside the other. The inner portion is almost transparent. Its development is arrested, and it never grows much longer than one centimetre, and often less. When it was first formed by splitting of the rudiment it possessed all its cells each of which increased in size until the maximum was attained, in the process losing the chlorophyll and protoplasm. When fully grown it becomes reddish-brown and membranous, consisting of dead cells and serving as a scale. It has a function, and that is to protect the bud-leaves next inside, and these it completely enfolds.

The development of these scales suggests the formation of connate stipules, and in an adult stem they are visible as small scales in the leaf-axil closely adpressed to the stem. Their function is done.

Transverse section of a young leaf in the bud.

The outer prismatic half of the leaf-rudiment develops into the foliage-leaf. Hairs develop on both sides in large numbers (Fig. 4), while the leaf is still prismatic and while all its cells are actively dividing. All these hairs, however, die before the leaf is fully expanded. It has been suggested by Boodle and Fritzsch (Comparative Anatomy of Dicotyledons, p. 245) that since in the different species of Rhus the hairs are variously shaped, they might furnish

useful characters for specific distinction. According to them, among the many species investigated, two alone, R. acuminata, D.C., and R. semialata, Murray, possess papillae on the lower surface of the leaf. As regards the genus Corynocarpus no mention is made of any hairs. In Corynocarpus the hairs vary among themselves. As has been stated, they occur on the adaxial surface of the scale, and on both surfaces of the leaf-rudiment, early in development; also they vary in shape (Fig. 5, b, c, d). On the scale they take the form of double rows of cells with pointed or blunt ends (Fig. 5, e), or sometimes they may be club-shaped outgrowths. On the leaf itself they are in almost every case globular. They occur in large numbers but are never so crowded as on the leaf-base. When first formed, in

A Vertical Section through a hair.

proportion to the size of the little leaf they are very large, and their development is completed long before the leaf unfolds. The uniseriate outgrowths and unicellular hairs found in some species belonging to the family do not occur in the karaka. The dense crowding of hairs on the scale may in part account for their being elongated; it may possibly also, in the first place, influence the direction of the dividing walls. In both leaf and scale a hair arises from one (or sometimes two) epidermal cells. The cell (or cells) becomes slightly larger than its neighbours, then divides in planes almost perpendicular to the leaf-surface, so that a small group of cells is formed which projects slightly above the epidermis. The direction of the first dividing walls is not constant; it may vary from

almost perpendicular to very oblique, also the number of epidermal cells that take part in the formation of a hair may be more than two. The next dividing walls are at right angles to the first, and a small rounded outgrowth on the surface of the leaf is the result (Fig. 5, a). Further divisions take place irregularly as the hairs increase in size, so that the hairs are not all the same shape (Fig. 5, b, c, d.).

While a globular hair is developing a disturbance takes place in the hypodermal tissue. The cells surrounding the hair and at its base become drawn up into a convex shape, and it would seem that they enter into and take part in the formation of the hair. However, this is only apparent. The hypodermal cells take no part in the development of hairs.

A description of the leaf-anatomy is given by Hemsley (Annals of Botany, September, 1903), but he makes no mention of any hairs. He describes a 1 to 2-layered hypoderm just beneath the epidermis, and mentions that “cork-warts occur in small numbers on the lower epidermis.” It seems that these cork-warts are formed by dead hairs, for as soon as the hairs are fully grown they begin to die off, first becoming golden brown. By this time the leaves begin to assume their characteristic dark green colour. The cells of the hairs collapse, as do also the cells of the convex hypodermal group, and the cells of the epidermis lying above them. This causes the outline of the epidermis to fall in just at that point and the hairs come to lie each in a depression, whose entrance they block (Fig. 6). In

A depression on the leaf surface.

many cases the hairs drop out altogether, leaving minute holes like pin-pricks. The layers of cells forming the floor of a pit were once the hypodermal cells, and they are arranged very regularly, after the fashion of cork cells (seen in section). They consist of either cutin or suberin and together with the cuticle stain the same shade of yellow with chlor-zinc-iodine, and the same different yellow with potash. They are more or less soluble in concentrated sulphuric acid. The dead hairs form the cork-warts. This formation of pits applies to the leaf alone, for on the leaf-scales the formation of hairs causes no convex swellings or subsequent depressions. When the time comes the hairs die and drop off, leaving no mark.

A marked peculiarity is found in the epidermal cells surrounding a hair. If strips are taken from the epidermis, these cells are seen to be arranged concentrically and radially, and there are very many concentric circles of cells. This arrangement of cells is so obvious in a surface view that it is difficult to understand why it should have been so long overlooked. In this concentric arrangement the hypodermal cells take no part (Fig. 7).

Surface view of a leaf depression.

The hairs seem, therefore, to have no special function. In the leaf-scale they may be slightly protective, for it is only there that they are in any way a covering, but even on the scale they are not present except in the bud, and on the leaves they are completely protected by the leaf-scales immediately outside them. It was not determined whether the hairs are absorbent or otherwise. Very thick sections possessing whole hairs were taken, and placed in eosin. They were left in some cases for three days, but at the end of the time, although the sections were deeply stained everywhere else, the hairs remained colourless, except that in one or two cases the eosin entered a hair from its base.

Hemsley notes the large numbers of clustered crystals in the plant. They are calcium oxalate and occur in all parts, even in the petals. Schimper says that calcium oxalate formation in leaves is connected with the appearance of old age, but in the karaka the crystals are abundant in even the youngest leaves, also in all parts of the flower.

Flower.—The flower-buds are very slow in development, so that from the time the buds first appear, in June and July, it is some months before the flowers open, in late September and October. In very small buds, 1 mm. in diameter, the ovary is fully formed and the carpels have closed in. The buds increase to a diameter of 3 mm. and then open; the flowers are 5 mm. across and pale green. The style in buds of 1 mm. diameter is very short, the stigma colour-

less, and bent at right angles to the style. The anthers are as tall as the gynoecium with very short thick filaments. The stigma slowly develops a bright red colouring matter while still in the bud, but as soon as the flowers open the red pales to yellow. The anthers dehisce almost as soon as the flowers open. The style remains always short with the stigma bent. The ovary is one-celled and swollen on one side, and on this side the ovule is inserted.

Pollination.—The five anthers are introrse, and the stigma exposes its stigmatic surface about the same time as the anthers dehisce. Bees and blowflies visit the flowers when the days are not too windy, and probe right down to the base. However, it seems probable that the flowers are wind-pollinated, for compared with the enormous number of flowers produced, the amount of seed set is very small, and it is not at all an unlikely thing to find no fruit on trees which were covered with blossom. Pollen-grains lodged on the stigma germinate in large numbers almost immediately, penetrating the

The style is split, showing pollen tubes.

style like hyphae, almost to the base (Fig. 8), but some weeks seem to elapse before fertilization takes place. This may partly be the reason why seed so often fails to set, and the prevalent gales at this time of the year, November, which blow almost all the flowers from the trees may account for the rest. If a flower is allowed to remain on the trees for some time after pollination but before the pollen tubes reach the ovary the pedicel withers and the flowers snap off at a touch.

Discs.—Each disc consists of a swollen yellow portion united at the base to a petaloid process. In development the petaloid process forms first, the disc appearing later as a minute elevation on the inner surface. A vascular strand enters from the base of the flower and passes into the petaloid process, bending slightly in the direction of the disc, but not entering it. The cells of the petaloid process soon cease to divide and growth becomes limited, but the cells of the swollen portions remain meristematic for a longer time, dividing very rapidly until full size is attained. The cells are typical meristematic cells, very small, with dense protoplasm, all the cells dividing by ordinary cell-division as long as they are active. Although Cheeseman calls these lobes staminodia they never, so far as these observations went, show any sign of becoming staminal structures. There is never any indication of an archesporial cell or cells. It would seem that the petaloid processes appear to represent an inner series of petals, each bearing on its inner surface a swollen portion

bright yellow in colour, which is possibly glandular. The insects which visit the flowers probe about at the base of the petals just where the yellow lobes are situated, but there is nothing in the structure of the lobes to show that they secrete any substance. The cells always remain typical meristem cells until they cease to divide, and by the time the flowers are wide open and insects are paying their visits they become almost empty of cell contents. The other organs of the flower mature at the same time, live for a short time, then they all die.

A Vertical Section through a young Flower Bud. s = sterile carpel; f = fertile carpel; an = anther.

Carpels.—In the formation of the pistil two carpels take part (Fig. 9). Two fibro-vascular strands from the pedicel enter the carpels separately, and the two carpels are unlike, only one being fertile. When mature the fertile sporophyll is much reduced (Fig. 10), being one half the length of the sterile sporophyll, and it bears a single pendulous ovule from its upper end. It takes no part in

the formation of the style, its function being to produce the ovule, and it contributes less than half towards the formation of the ovary. The sterile sporophyll forms the rest of the ovary and the whole of the style. In the development and formation of the pistil the two

sporophylls meet, so that the margins of the one touch the margins of the other, but the pistil is not radially symmetrical but is bilateral (Fig. 11), first, owing to the unequal length of the two sporophylls, and, second, because the short sporophyll is rounded to contain the

OV. = Ovary; O = Ovule.

ovule. The ovule begins to form very early, when the two sporophylls are almost the same size, and before the margins grow together (Figs. 9 and 12). To form the style the upper half of the sterile sporophyll folds inwards and the margins meet; however, this fusion is not complete, for there is always a groove left running the length of the style. There is also incomplete fusion between the upper part of the fertile sporophyll and the base of the style, so that an opening is left into the ovary at the base of the style, and this opening does not close up until after fertilization. The stigma remains bent (Fig. 11).

Ovule.—The ovule begins as a bulge on the surface of the fertile sporphyll very early in its life (Fig. 12). A group of cells just

Young Ovule.
Young Anther.

below the epidermis all divide very rapidly in different directions until a pendent group of cells is formed, the nucellus. This rudimentary ovule grows to a diameter of .04 mm. before the integuments begin to form, and this stage is found in unopened buds. This is the period of most rapid growth of the funicle, only a very small

proportion of the mass of cells constituting the nucellus at the apex; the nucellus remains in this small proportion until the archesporium is formed, when the buds are beginning to open. The integuments begin to develop almost at the same time.

A cell in the middle line of the nucellus, about the third or fourth cell in from the outside, becomes the archesporium (Fig. 14). It divides into two, an inner and an outer cell, and both cells are very large (Fig. 15). The outer cell divides again in a different direc-

a = Archesporium.
oi = Outer Integument.
i = Inner Integument.

tion also the cells between it and the outside of the nucellus divide several times, so as to increase the number of cells between the outer layer and the archesporium. The archesporium thus comes to lie at a depth of seven or eight cells in from the outside. Without further division the archesporial cell develops into the embryo sac (Fig. 16). The tiers of cells above it separate slightly in young ovules, leaving a passage down to the embryo sac (Fig. 17, a, b). This would seem to suggest a preparation for the entrance of the pollen-tube, were it not that the passage is closed later on. The anthers complete their development much earlier, and it is usual in July to find pollen-sacs containing cells in tetrad

division. Further development of the ovule takes place most markedly in the nucellus and integments, for as soon as the arche-

Different stages in development of Embryo Sac.

sporial, cell is clearly differentiated the funicle almost ceases to grow. Growth is more rapid on one side, the ovule begins to curve and finally becomes anatropus. The embryo-sac is oval and becomes

Only two Antipodal Cells are shown.

enlarged at the expense of the nucellus. When the length across is from about .002 mm. to .06 mm. the contents are arranged in typical fashion, consisting of two synergidae, an egg-cell, a well-marked secondary nucleus, and a varying number of antipodal cells (Fig.

p.n. = Polar Nuclei.
m. = Microyple.
i. = Inner Integument.
o.i. = Outer Integument.
n. = Nucellus.
e.a. = Egg Apparatus.
Ant. = Antipodal Cells.

17, c.); there may be as many as seven or eight (Fig. 18). Further increase in size of the embryo-sac does not take place until pollen-grains have germinated on the stigma.

Stigma.—In unopened buds the stigmatic surface consists of a layer of very large evenly arranged cells, which are in longitudinal section more or less square in outline. When the stigmatic surface is ready to receive pollen these cells separate from one another and

stand a little apart (Fig. 19, a, b). The cells immediately beneath these are small square cells which stain readily, and it is these cells that in unopened buds contain the red colouring matter.

Pollination takes place in November and December, and although a very obvious groove is left along the style by the infolding of the sterile carpel, pollen-grains do not grow down the groove. They always penetrate the tissue of the style, gaining nutriment, and germinate in large numbers. When they come to the opening left by the incomplete fusion of the two carpels, they make their way down

st. = Stigma.

it and enter the ovary near the base of the funicle. Two or three pollen tubes may enter the ovary, one tube is seen to creep over the surface of the ovule near the micropyle, and only one seems to find

its way in (Fig. 20, a, b.). It finds its way along a little space left between the apex of the nucellus and the integuments, at the side. Actual entry into the embryo-sac was not observed, although in many cases an egg and a sperm nucleus can be seen lying side by side in the same envelope (Fig. 21). Double fertilization has not been observed, but at the time that the egg and sperm nuclei are lying side by side, other cells are formed very rapidly at the micro-

pylar end of the sac. Some very clear cases were observed where these nuclei are all dividing before any appearance of fertilization. These cells newly formed have abundant protoplasm and are at first grouped closely together, but they are not provided with cell walls until later. They soon come to line the embryo-sac, still dividing very rapidly and often are all found in telophase, showing spindles. Then their protoplasm becomes thinner and they secrete cell walls. Embryo-sacs presenting this state of affairs vary in length from .3 cm. to .6 cm. It would seem that these cells are prothallial cells formed before fertilization is complete, distinctly a gymnospermous character.

Embryo.—The embryo divides across, forming two cells (Fig 22), a basal cell and a terminal cell. Of these the basal cell does not divide again, and the terminal cell, unlike that in typical Angiosperms, forms the embryo. It divides across twice and further divisions result in a spherical embryo of about sixty-four cells or