Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 67, 1938
This text is also available in PDF
(5 MB) Opens in new window
– 149 –

Ovule Anatomy and Development and Embryogeny in Phyllocladus alpinus (Hook.) and in P. glaucus (Carr).

[Read before the Otago Branch, May 11, 1937; received by the Editor, May 12, 1937; issued separately, September, 1937.]

Contents.

1.

Introduction and Historical.

2.

Methods and Material.

3.

Anatomy of the strobilus, vascular supply and general morphology of the ovule.

4.

Development series:

(a)

Embryosac:

(b)

Endosperm;

(c)

Archegonia and fertilisation. P. glaucus.

(d)

Embryogeny.

5.

General discussion.

6.

Summary of affinities of Phyllocladus.

7.

General summary.

8.

Bibliography.

Introduction and Historical.

Phyllocladus (L. C. Rich), a genus of shrubby conifers distinguished by the complete loss of leaves except in the seedling plant, and having the branchlets flattened into broad phylloclades, contains seven species of which three, P. trichomanoides (D. Don), P. glaucus (Carr), and P. alpinus (Hook.), are found in New Zealand. Another Species, P. rhomboidalis (Rich), is found in Tasmania, P. hypophyllous (Hook.) in Borneo, P. protractus (Warb) in montane forests in the Philippine Islands and P. major (Pilger) in New Guinea. P. alpinus closely resembles the Tasmanian P. rhomboidalis, differing according to Cheeseman (1925) principally in the position of the ovules. Kirk (1877) considered that the two species should be amalganated. The same writer (1889) also reports the occurrence of P. rhomboidalis in New Caledonia. the distribution of the species within New Zealand is as follows:—

P. trichomanoides is found in the northern wet forests from N. Cape to Taranaki. In the South Island it is restricted to the N.W. Botanical district. Altitudinal range 0–2500ft.

P. glaucus is confined to the north of the North Island. 0–2000ft.

P. alpinus occurs in montane and sub-alpine forests in both islands. In the south of the South Island it descends to sea-level. 0–6000ft.

The genus has been placed by various writers in the families Taxineae and Podocarpineae and has also been regarded as constituting a separate intermediate family, the Phyllocladaceae. Its resemblances to the Podoarps, however, proved too strong to justify the

– 150 –

retention of this intermediate family, and Pilger's (1926) classification of it as sub-family III, Phyllocladoideae, of the family Podocarpineae, seems the most satisfactory. Pilger subdivides the genus into Section A, P. glaucus and P. trichomanoides, and Section B in which the remaining species are placed. Taxonomic accounts of several of the species have been published, but all knowledge of the morphology of the ovule has been obtained from studies of P. alpinus. One of the earliest accounts was given by Kirk in 1877, but he dealt solely with the external morphology and classification of the New Zealand species. The next paper of importance was published by Dr. Agnes Robertson (1906) in the Annals of Botany (1906). She dealt with the vascular anatomy of the stem and of the ovule, but did not obtain any developmental stages. These were first described by Miss Kildahl (1908a) and Miss Young (1910). Both of these authors were able to give a full series of stages in the development of the pollen-grain and of the male gametophyte, and Miss Young obtained a fairly complete series of archegonia. Nothing was known of the embryogeny save that both authors reported an eight-nucleate stage before wall-formation in the proembryo and the presence of two cotyledous in the mature embryo. All this work was done in meterial of P. alpinus obtained through Dr. L. Cockayne. For the Tasmanian species P. rhomboidalis a sketchy account of the main anatomical details was given by Baker and Smith (1910) in their monograph entitled The Pines of Australia. I have seen no account of the remaining species.

The purpose of the present research was to obtain a more complete series of stages in the development both of the strobilus and of the ovule in P. alpinus, to investigate its embryogeny and to commence a study, along similar lines, of another species, for which purpose P. glaucus, a member of Section A, was chosen.

Methods and Material.

The material for this research consisted of ovulate strobili of P. alpinus collected as follows:—

1.

18 October, 1935. Dunedin Botanic Gardens.

2.

6 November, 1935. Dunedin Botanic Gardens.

3.

23 November, 1935. University Museum Gardens.

4.

11 December, 1924. University Museum Gardens.

5.

12 December, 1935. University Museum Gardens.

6.

13 December, 1933. University Museum Gardens.

7.

26 December, 1935. Dunedin Botanic Gardens.

8.

29 December, 1935. Dunedin Botanic Gardens.

9.

29 December, 1935. University Museum Gardens.

10.

21 February, 1935. Dunedin Botanic Gardens.

11.

22 February, 1936. University Museum Gardens.

12.

13 March, 1936. Dunedin Botanic Gardens.

13.

January, 1934. Colac Bay, Southland.

The material collected in the University Museum Gardens came from a single isolated tree and it was noted on making the collections that a certain number of the strobili showed whitening and shrivelling. When these were examined the ovules were found to have

– 151 –

aborted when in the free nuclear condition of the embryo sac, the abortion being accompanied by general disintegration of the nucellus. Fifty or more complete strobili from collections 3, 4, 5 and 6 were microtomed without finding a fertile ovule, but also without finding anything to account for the abortion except that no trace of pollengrains were seen. From field experience in the sub-alpine habitat of P. alpinus, I knew that very rarely was a shrub seen with a good yield of fertile seeds. With this in view a large collection of seeds from Colac Bay was examined. In the thirty apparently fertile ovules dissected out no trace of endosperm or embryo was found.From collection 9 a few ovules with well-developed endosperm were obtained but these also did not contain any embryo. On the other hand all ovules collected at the Dunedin Botanic Gardens were fertile.

This wholesale abortion is probably due to lack of pollination. According to Cheeseman (1925) P. alpinus is usually monoecious, but apparently dioecism is much more widespread than was believed. Why the abortion should occur when the embryo sac is in a free nuclear condition, i.e., at the usual time of pollination, would have to be explained on the grounds of some philosophic “loss of a stimulus consequent on pollination.”

As well as the above material of P. alpinus a certain number of collections of ovulate strobili of P. glaucus were made:

1.

29 December, 1935. Dunedin Botanic Gardens.

2.

12 January, 1936. Dunedin Botanic Gardens.

3.

22 February, 1936. Dunedin Botanic Gardens.

4.

13 March, 1936. Dunedin Botanic Gardens.

Again, according to Cheeseman (1925), P. glaucus may be either monoecious or dioecious. Thus, when ovules collected on January 12 were found to be in various stages of fertilization it was assumed that the tree, an isolated one, was monoecious, and collections 3 and 4 were made. Of each of these some dozens of ovules were examined, but no later stage than that of the actual fusion of the nuclei was present. This seems to be due to the tree being dioecious, the pollengrains observed coming from a neighbouring plant of P. alpinus. These facts reduced the number of useful collections to Nos. 1, 2, 7, 8, 10, 11 and 12 for P. alpinus and to Nos. 1 and 2 for P. glaucus.

All material was killed and fixed immediately on collecting in Carnoy's Fluid (acetic acid, chloroform, 95% alcohol) in vacuo. The material was then preserved in 95% alcohol, with, for the older stages, an addition of glycerine. For all stages prior to the formation of endosperm the whole strobilus was blocked in paraffin and cut, but for later stages the ovules could be dissected out. As soon as the stony middle integument had developed it was found difficult to infiltrate the objects, but this was overcome by removing the tips of the ovules and leaving in the paraffin bath for a week or more. This prolonged exposure to heat (51–5° C.) did not seem to result in any damage to the ovules.

Objects were cut with a Cambridge Rocker microtome, but this was found to be incapable of dealing with the stony integument or with the basal palisade. Better results were obtained with a heavy

– 152 –

sledge microtome, but cells of the integument continued to pull loose and cut across the face of the object. To remedy this several methods were tried, as follows:

1. After blocking out the object the cut tip of the integument was bared and the whole object was immersed in water for a week or more (Couch, 1928). This effected an improvement, but not to a satisfactory extent.

2. Using the same technique but with alcohol replacing water (Rawlins, 1933). This did not seem to make any difference.

3. By the use of hydrofluoric acid (Langdon, 1920). Many strengths were tried, 10%, 20%, 35%, 50%, of the pure and of the commercial acid. Objects were left in for periods of from a week to a month. The 10% acid caused slight plasmolysis but did not affect the stony integument. The 50% acid even for one week resulted in very great plasmolysis and caused complete precipitation of nucleoprotein, in many cases leaving only the nuclear membrane intact. The 20% and 35% acids, while not bringing about satisfactory softening, did not damage even the archegonia to any great extent. Slight plasmolysis of all cells took place, but the most marked result was in staining. After treatment with the acid the cell-walls did not stain with Haidenhain's Iron-alum Haematoxylin although the middle lamellae could be plainly distinguished. No great difference existed between leaving the objects for one week or two weeks in these strengths, but after one month they were completely ruined. After treatment with hydrofluoric acid very great care had to be taken to ensure its complete removal. If this was not done fading of the stain was very rapid.

4. An attempt was made to remove the integuments, but owing to the small size of the nucellus such an undertaking always resulted in damage to the contents of the embryo sac. After the ovule had been blocked out, however, it was found comparatively easy to remove the integuments using the microtome with an old knife. If the ovule had been previously treated with hydrofluoric acid, or had, since being blocked out, been treated as in methods (1) and (2) the integuments split off more easily still. Once the integuments were removed the objects were returned to the paraffin bath and reblocked.

Sections were cut at 5, 7, or 10 μ, the only consideration being the thickness at which the best ribbon was obtainable on that particular day. Several stains were tried. For certain features in the vascular anatomy of the strobilus safranin and light green were used, but these gave patchy results when used for the details of the embryo sac. Safranin and gentian-violet were also tried without much better results. A few slides were stained with Delafield's Haematoxylin, but 95% of the staining was done with Haidenhain's Iron-alum Haematoxylin. At first iron-alum was used both as a mordant and for destaining in the orthodox way, but later destaining was carried out with picric acid (Tuan H'su Chuan, 1930). Various staining schedules were experimented with the following gave most satisfactory results:

I.

Paraffin-xylol-alcohol-water.

II.

Iron alum 4% for 1 hour. Wash.

– 153 –
III.

Haematoxylin, filtered, 12 hours. Wash.

IV.

Saturated solution of picric acid. Destain until the sections are a light brown. Approximate length of time may be judged from the depth of colour reached in washing after III.

V.

Tap water for 2 hours. The longer the time in water, within limits, the bluer and more brilliant the stain. Scott's Tapwater Substitute may be used to shorten the time. Pass up the series as in I and mount.

This schedule gave excellent nuclear staining as well as showing details of the cytoplasm and of the cell-wall. Lignified walls such as those of a stony integument, the basal palisade and tracheids stained bright yellow with the picric acid. The cuticle of the outer integument stained orange showing clear zonation, while the megaspore membrane took on a rich brown colour. The schedule has also been used with excellent results for staining young flower-buds and anthers of Lilium martagon.

Drawings were made for the most part with a Zeiss reflex apparatus, but an Abbe camera lucida was also used. An attempt has been made in the plates to keep all items on a comparable scale and to maintain the clear relations of the parts by means of habit drawings.

Anatomy of the Strobilus, Vascular Supply and General Morphology of the Ovule of P. alpinus.

The most complete account of the strobilar anatomy of Phyllocladus alpinus was given by Miss Robertson (1906). The strobili are borne laterally near the base of the phylloclades, 1 to 5 being found on each phylloclade. Each strobilus consists of six to eight fleshy scales, the majority of which have a single erect ovule in their axil. Fig. 1 shows a longitudinal section of a young strobilus in which two ovules in free nucleate embryo-sac condition are seen. Vascular strands run out into each bract and inversely orientated strands to the base of each ovule, where they terminate in a tracheidal plate. There are numerous oil-ducts containing an oil-ducts containing an oil of high terpenic content and large air-spaces. At the base of each ovule an aril develops. It originates immediately outside the tracheidal plate and grows rapidly, remaining two or three cells thick and often developing a hollow structure. Stages in aril formation are seen in Figs. 2, 6. The mature aril persists as a light papery cup extending two-thirds of the way up the ovule (Figs. 7 and 8). The interpretation of this aril or epimatium is dealt with in the general discussion.

Two layers only can be distinguished in the integument of a young ovule—the outer fleshy, two layers thick, and the inner consisting of numerous smaller cells. The middle stony layer develops first near the micropyle about the time of pollination, i.e., when the embryo sac is in a free nuclear condition, and, by the time the archegonia have formed, has extended to the base of the ovule. Thickening of the radial walls takes place in cells immediately under the outer fleshy layer, which at the micropyle may be only one cell layer thick. The thickening later extends to the inner tangential

– 154 –

walls, but the outer walls remain thin even in the mature seed. Except near the micropyle the stony layer remains one cell layer thick. Numerous pits are left in the walls and each cell has a large nucleus which lies close to the thin outer wall, and a conspicuous crystal. Tests with hydrofluoric, hydrochloric and sulphuric acids failed to dissolve these crystals. The outer fleshy layer, protected on the outer side by a three-zoned cuticle, persists until the seed matures. Fig. 3 shows the first cell of the stony layer, and Fig. 5 and 4 show the stony layer in longitudinal and transverse section respectively.

The vascular supply to the ovule is of considerable interest. From the main axis of the strobilus two inversely orientated bundles run out to the base of the ovule. Here they end in a tracheidal plate. In a very young ovule both this plate and the basal palisade above it are represented by a zone of cells with conspicuously large nuclei. The aril develops around the outer edge of this zone (Fig. 2). The cells of the tracheidal plate differentiate as short tracheids with marked “Taxinean” sculpturing, that is, they have both spiral and annular thickenings as well as bordered pits (Fig. 11). Immediately above the tracheidal plate a zone of elongated cells develops in much the same manner as did the cells of the middle integument. Thickening commences on the radial walls and later extends to the end walls. Numerous pits are left and each cell retains its nucleus and has a characteristic large crystal (Fig. 9).

At the outer margin the palisade interlocks with the stony layer of the integument (Fig. 9), but the palisade does not completely shut off the ovule from the tracheidal plate. Two opposed gaps are left, into each of which several tracheids run (Fig. 10). These gaps doubtless represent the position of the original vascular supply to the inner integument. It is possible that the palisade layer and the stony layer of the integument are, by their possession of numerous pits and thin end walls efficient in the translocation of materials.

Throughout the vascular system of the strobilus centripetal xylem is found—in the main strobilar axis, in the strands supplying the fleshy scales and in the vascular strands supplying the ovule. Fig. 12 shows a transverse section of one of these latter strands. The nature of the centripetally formed tracheids has been verified by longitudinal sections.

The young ovule (Fig. 1) has a wide micropyle. After pollination has been effected the micropyle is closed by the development of several layers of the stony integument and also by divisions taking place in the cells of the inner integument. The nucellus is free to the base, and, unlike Torreya and Taxus, remains free. The relative thickness of the various integuments of the air-space and of the nucellus of an almost mature ovule is shown in Figs. 7 and 8, which are respectively longitudinal and transverse sections of ovules at a time of embryo development corresponding to Figs. 43 and 44. In the mature seed the nucellus is represented simply by a membrane.

Young stages in the development of the nucellus have not been obtained, but collections of these are now being made.

– 155 –

Developmental Series.

The details embodied in this section refer to prepared material of P. alpinus with the exception of those portions describing the development of the endosperm tissue and of the archegonia, which refer to P. glaucus. These are embodied in the text in order to add continuity to the description and also in view of the fact that archegonial development in P. alpinus, as described by Miss Young (1910), does not differ from that now described for P. glaucus. For the sake of clarity all references and discussion have been held over until the next section as far as possible.

(a) Development of the Embryo Sac (P. alpinus).

The embryo sac has been described as developing from the innermost of a row of three cells. It seems probable that a normal linear tetrad will be found. This stage is reached at the end of September, and by the second week in October the megaspore has enlarged considerably. Fig. 15 shows it in the 4-nucleate condition, although only two of the nuclei occur in the section drawn. The megaspore even at this stage possesses a definite megaspore membrane which is most noticeable on plasmolysis. It is surrounded by a layer of large glandular vacuolate cells, richly stocked with food and occasionally binucleate. These cells are tapetal in function and encroach on the neighbouring nucellar tissue. The megaspore usually shows slight plasmolysis, and it is believed that this is a natural condition rather than that it has been caused during killing and fixing, as the megaspore even when filled with endosperm is almost invariably in contact with the nucellus only in certain places.

At no time do the free nuclei more than line the periphery of the megaspore. This stage is shown in longitudinal section in Figs. 13–17. Fig. 14 is a transverse section of the megaspore and nucellus at the same age. Pollination takes places about this time, and immediately prior to the shedding of the pollen the cells at the apex of the nucellus undergo slight “disintegration.” This is evident in the comparison of the apex of the nucellus in Fig. 13 and that in Fig. 18. This modification would certainly seem to possess some slight “survival value” in presenting a more roughened surface for the lodgment of the pollen-grains. A similar occurrence has been described for Sciadopitys (Lawson, 1910), where a pollen cushion of large thinwalled cells at the tip of the nucellus, is differentiated. Again, this phenomenon would seem to be connected with the abortion of ovules at this stage in their development, consequent on non-pollination.

(b) Development of the Endosperm.

The endosperm tissue develops very rapidly. At the end of November the number of free nuclei lining the periphery of the embryo sac totals, as the average of several approximate counts, 256. In the study of P. alpinus no stage indicating the precise method of formation of the endosperm tissue was obtained, although the appearance of the endosperm at a later stage suggested, from its extraordinarily regular radial construction, something perhaps akin to that described for Gingko (Carothers, 1917, or Chamberlain, 1906). Later, however, the missing stages were obtained in P. glaucus (Fig. 19).

– 156 –

Wall-formation sets in and large radial cells, occasionally bi-or tri-nucleate, and reaching to the centre, are formed. These are not laid down on the megaspore membrane, but have a separate external wall of cellulose. It could not be determined whether or not this cellulose wall was laid down prior to, and independently of, cell-wall formation, and hence its analogy with the endosperm membrane as described for Gingko (Carothers, 1907) should not be stressed. Nuclear division of each primary radial endosperm cell sets in except in two cells at the apex of the endosperm (Fig. 20). These then become recognizable as the archegonial initials. Any earlier setting apart of them could only occur in the free nuclear stage. This tendency towards the earlier establishment of the archegonial initials has been remarked both in Pinaceae and in Taxaceae in many instances.

Each of the original primary radial cells of the endosperm has its own complete wall, but when the endosperm is fully developed the individual cells become crowded together and approximate to a normal tissue. The endosperm, however, retains its radial structure until late in the development of the embryo. The megaspore membrane continues to thicken and becomes differentiated into the normal two layers—the inner homogeneous and the outer layer of club-shaped rods (Figs. 52, 53). In the areas where the endosperm is in contact with the nucellus the phenomenon of “rumination” occurs. That is to say, the endosperm develops unevenly, giving the appearance of plates of perisperm invading the endosperm. This is not nearly so marked as in Torreya taxifolia (Coulter and Land, 1905). The endosperm later becomes multinucleate and richly stocked with oil globules. As many as nine nuclei may be found in a single cell except in those near the priphery, which are markedly smaller than the rest (Figs. 52, 53).

(c) Archegonia and Fertilisation.

The following description is taken from P. glaucus and does not differ in essential outline from that already described by Miss Young (1910) for P. alpinus.

The archegonial initials appear before periclinal segmentation of the endosperm has reached to the centre. The archegonia do not arise from the very great enlargement of single small superficial cells, but rather, consequent on the method of formation of the endosperm tissue, first appear as relatively large cells. Thus, in Fig. 19, any one of the primary radial cells at the apex of the endosperm might be visualised as a potential archegonium. It is only when the surrounding tissue has undergone much subdivision that the archegonial initials become recognisable (Fig 20).

The cells of the jacket-layer are at first indistinguishable from the cells of the surrounding endosperm (Figs. 20, 21). Later they undergo successive periclinal divisions to give rise to a much more compact tissue (Fig. 22). At this stage their multinucleate condition first appears. This results from continued nuclear division without subsequent wall-formation. The jacket cells elongate considerably and become densely packed with cytoplasm with numerous food

– 157 –

vacuoles. The cells of the mature jacket (Figs. 27–31) may possess as many as eight or nine nuclei, but two to five are the usual numbers. Miss Young (loc. cit.) described a case of two archegonia in a common jacket in P. alpinus. A possible case of this was found in P. glaucus, but as the section had to be interpreted with the help of Miss Young's figure, it would not stand alone as evidence for the existence of an archegonial complex. In any case I do not consider that the occasional occurrence of two archegonia in a common jacket in Phyllocladus bears comparison with the occurrence of archegonial complexes in other groups. It seems, rather, to depend solely on the casual position of the archegonial initials. Many cases were found in which two archegonia had a common jacket-layer between them (Fig. 25).

The archegonial initial divides periclinally, giving rise to a neck and a central cell (Fig. 20). The neck-cell undergoes two successive divisions to form a single plate of four cells (Fig 21). It seems probable that further subdivision of the neck occurs frequently. Miss Young (loc. cit.) described a case in which each of the neck-cells had divided, but without further wall-formation, the two cells of each pair being divided by a “Hautschicht,” or constriction of the cytoplasm. Fig. 26 shows two archegonial necks in transverse section, one showing seven and the other six neck-cells. In the latter a “Hautschicht” appears in two of the cells, bringing the number to eight. In no case, however, has any irregularity resulting in the production of a neck of greater thickness than a single cell layer been observed out of a total of archegonia examined amounting to over one hundred.

Miss Kildahl (1908a) had noted a peculiarity in the structure of the neck of mature archegonia of P. alpinus. This was explained by Miss Young (loc. cit.) as brought about by the extension of adjacent jacket-cells to form a membrane due to the rapid centrifugal growth of tissues at the archegonial end of the endosperm. This is found to occur also in P. glaucus. In the young archegonium the megaspore membrane is usually tightly adpressed to the outer surface of the neck-cells. The surrounding endosperm grows forward with the result that the archegonia are left at the bottom of an archegonial “chamber” and the megaspore membrane has been pushed clear (Fig. 22). During the growth of the endosperm the width of the archegonium has been increasing considerably. This is met by the stretching of the adjacent jacket-cells (Figs. 23 and 24). Thus, at the time of fertilisation, the neck appears in longitudinal section as a small group of two or three cells suspended over the top of the archegonium by a fine membrane (Figs. 27, 28, 29). Notwithstanding the apparent thinness of this membrane it offers considerable resistance to the pollen-tube (Fig. 28), and fertilisation always takes place through the neck itself.

Contemporaneously with the division of the neck-cell to give a plate of four cells, the nucleus of the central cell divides to form the egg-nucleus and the ventral-canal-nucleus, but no cell-wall is formed. The ventral-canal-nucleus is small and lies immediately under the neck (Fig. 25) and by the time of fertilisation has usually disappeared. The egg-nucleus at this stage is also small and lies

– 158 –

at the top end of the archegonium. It increases in size very rapidly attaining an ultimate diameter of from one-half to two-thirds of the width of the archegonium (Fig. 29). The whole archegonium becomes densely stocked with cytoplasm.

For the sake of comparison a mature archegonium of P. alpinus is shown in Fig. 30. Fig. 32 is a habit drawing of the apex of the nucellus and endosperm at about the same age.

Various stages in fertilisation were observed both in P. glaucus and in P. alpinus, but the former must be interpreted with caution as the ovules were obtained from an isolated female plant and it is believed that the pollen originated from a neighbouring shrub of P. alpinus. Although a fusion-nucleus was formed, no embryo developed, as was seen from a large series of ovules. The endosperm, however, persisted and the integuments became fully developed.

Pollination, in P. alpinus, takes place during the first two weeks in November. The winged pollen-grains Iodge on top of the nucellus and there germinate. Complete stages have been traced both by Miss Kildahl and Miss Young. Figs. 32 and 35 show the passage of the pollen-tubes to the neck of the archegonia. They digest their way through the nucellus and pass easily through the megaspore membrane into the archegonial “chamber.” At this stage four nuclei are present in the tube—the stalk, the tube, the body-cell and the secondary prothallial nuclei. Fertilisation in P. alpinus takes place during the first week in December. In P. glaucus it takes place, in Dunedin, about three weeks later.

Immediately prior to the discharge of the pollen-tube contents the body-cell nucleus divides unequally giving two male nuclei. The pollen-tube contents may be discharged through the neck of the archegonia or just outside. A certain amount of male cytoplasm accompanies each male nucleus (Figs. 27, 28) and this forms a finely granular sheath around the fusion-nucleus. This seems to have a physiological nutritive function. The second male nucleus and the tube, stalk, and prothallial nuclei disintegrate and add to the cytoplasm of the egg.

Stages in fertilisation are figured in Figs. 27 and 28 for P. glaucus and in Fig. 31 for P. alpinus. These are self-explanatory, but it should be noted that in Fig. 27 the second male nucleus is much larger than usual and in Fig. 28 the egg-nucleus is not in median section and hence does not appear in its correct relative size.

(d) Embryogeny in P. alpinus.

The development of the embryo proceeds immediately. Miss Kildahl (loc. cit.) figures the four and eight nucleate stages and Miss Young did not obtain a greater number of free nuclei than eight. It is evident that these stages are passed through very rapidly. In one gathering in the present study many young embryos in various stages of suspensor elongation, together with mature archegonia, fertilised and unfertilised, were obtained; but although some forty of these ovules were microtomed only two sections showed embryos in the free nuclear condition. These are shown in Figs. 33 and 34. Both have eight free nuclei, which are more or less paired and are situated near the base of the archegonia.

– 159 –

The nuclei migrate to the base of the egg and become arranged in three approximate tiers. Two constitute the embryo proper and the third elongates to form the suspensor. A young embryo immediately subsequent to wall-formation is seen in Fig. 35 situated at the bottom of an archegonium in which the second male nucleus is still visible. The cells of the suspensor elongate very rapidly and push the embryo well down into the endosperm. One of the cells at the apex of the embryo early becomes established as an apical cell and functions as such for some seven or eight successive divisions although not always with such extreme regularity as is the case in one embryo in Fig. 42. Apical cells may be traced in Figs. 36, 37, 39, 40 and 42. The cells behind the apex divide periclinally and anticlinally eventually giving rise to a spherical embryo such as those in Figs. 43, 44, and in which all trace of an apical cell has been lost.

From now on growth appears to take place in a series of stages—a period of cell enlargement followed by one of rapid division—in which all the cells behave with extreme regularity. For example, Fig. 45 is a transverse section of a slightly younger embryo than those in Figs. 43 and 44, but while the diameters of the embryos are approximately the same, the size and numbers of the cells are greatly different. In Fig. 46, a somewhat older embryo, all the cells appear to have undergone a period of rapid subdivision. This ovule had been treated with hydrofluoric acid, the cell-walls consequently not staining with Haidenhain's Iron-alum Haematoxylin, only the middle lamellae being visible. Marked plasmolysis had taken place, however, and this was much more evident between each group of cells than between the cells of each group. The same division of the embryo into groups of cells is visible in Fig. 47.

In the majority of ovules in the stage of the elongation of the suspensors, two embryos are present, as in Fig. 42. In Fig. 47 two embryos are shown, developing to an abnormally late stage. Usually one would have aborted long before this size is reached. No case of secondary embryos formed either by budding or splitting of the parent embryo have been noted (see general discussion).

Soon after the embryo has reached the size of that in Fig. 46 periclinal divisions set in near the apex of the embryo and anticlinal divisions at the centre. This results in the development of the two cotyledons and in the differentiation of the elongated cells of the central cylinder. The first of these periclinal and anticlinal divisions are shown in Fig. 48. Growth in size of the embryo now takes place evenly and rapidly. In Fig. 49 the first of the series of regular divisions, which are to differentiate the root apex, have taken place. Growth now takes place in length and by the time the seeds are shed at the beginning of March the embryo has reached the condition shown in Figs. 50 and 51. Fig. 51 is a median longitudinal section through the apex and between the cotyledons of an embryo which now occupies approximately one-fifth of the length of the endosperm and lies well at the centre of it. The embryo passes the winter in this condition, further development taking place in the spring.

– 160 –

General Discussion.

It is mainly on the evidence afforded by the pollen-grain and by the male gametophyte that Phyllocladus has been definitely placed in the Podocarpineae. Although no fresh work has been done on this side of the problem in the present research, it is as well to recapitulate briefly the main facts before going on to a discussion of the significance and affinities of certain features in the development of the ovule.

The male sporophylls bear two abaxial sporangia as in all podocarps. The pollen-grains are winged, this condition being also found in Podocarpus and in Dacrydium. In Microcachrys (Thomson, 1908) the wings are not so well developed and in Saxegothea (Stiles, 1908) they are absent. If, as Thomson (loc. cit.) suggests, this winged condition has arisen independently in Podocarpineae we are afforded a podocarp line along which Phyllocladus has developed. Throughout Podocarpineae the male gametophyte cuts off two prothallial cells as in Abietineae, but these later undergo a further division which is somewhat irregular. For example in Dacrydium (Young, 1907) in some species both cells divide, in other species only the second does so. This condition is repeated in Microcachrys (Thomson, loc. cit.) and in Saxegothea. In Podocarpus (Jeffery and Chrysler, 1907) there is much greater variation, one to eight prothallial cells being formed. In Phyllocladus (Kildahl, 1908a; Young, 1910) the first prothallial cell does not usually persist but the second may divide. This behaviour is in marked contrast to that shown by all members of the Taxineae, in which no trace of prothallial cells is found. The prothallial nuclei in most cases become free, and at pollination the tube usually contains several vegetative nuclei as well as the stalk, tube and body-cell nuclei. In spermatogenesis both in Taxineae and in Podocarpineae two unequal male nuclei are formed, and, although this is more marked in Taxineae, its occurrence in Phyllocladus is thus not exceptional.

The retention by the male nucleus of a complete sheath of male cytoplasm, which later envelops the fusion nucleus and appears to act in a physiologically nutritive capacity has also been observed in Torreya (Coulter and Land, 1905) and in Cephalotaxus (Coker, 1907), and is probably present in other members of both families.

Phyllocladus, to a very great extent, possesses primary centripetal xylem, but, notwithstanding the fact that mesarch structure is most common amongst Taxineae, this cannot be regarded even as indicating any relationship. The extreme reduction of the leaves to minute scales and the broadening of the secondary branchlets to phylloclades is itself a modification so profound, and would entail such deep-seated physiological changes, that extensive recapitulation of ancestral characters is not unusual. Similarly, “Taxinean” sculpturing of the tracheids, although characteristic of the Taxineae as a whole, is not by any means confined to them, and the presence of this type of tracheid in Phyllocladus is, by itself, of little phylogenetic significance. Of greater interest is the vascular supply to the ovule and the nature of its integuments.

In its erect free axillary ovule Phyllocladus resembles the Taxineae and differs from the Podocarpineae, but in this connection

– 161 –

it must be remembered that in the latter family all stages between the free erect condition and the inverted fused condition of Podocarpus are present. Also the asymmetry of the epimatium as compared with the arillus must be regarded in correlation with the degree of inversion of the ovule. There has been considerable confusion with regard to the interpretation of the arillus. In Torreya (Coulter and Land, 1905) the first integument to develop differentiates into two zones, an inner fleshy and an outer stony, and the second integument develops fleshy. It was natural to see here the three-layered testa of the older gymnosperms. This interpretation, that the arillus resulted merely by the retardation in development of the outer fleshy layer, could be applied also in the case of Taxus, except that it remained free from the stony layer; but in Phyllocladus the first integument itself gave rise to all three layers and it is evident that the aril is an entirely new structure. The absence of the outer fleshy layer in Torreya and Taxus can on this basis be explained as resulting directly from the development of the aril.

In Torreya and Taxus the inner integument and the nucellus are free in the young ovule, but interchalazal growth finally results in the nucellus remaining free only at the tip. This is not the case in Phyllocladus where the nucellus and inner integument remain free to the base. In Phyllocladus, in Microcachrys and in Saxegothea the vascular supply ends at the base of the ovule, whereas in Torreya (Oliver, 1903) there are two opposed vascular strands in the outer integument as well as indications of a nucellar vascular supply. The only indication of an ancestral condition in Phyllocladus is in the presence of two opposed gaps in the basal palisade, into each of which several tracheids from the tracheal plate run. These apparently represent the former vascular supply to the inner integument. The exact nature of the cells of the basal palisade has not been determined. Transition stages between them and the cells of the tracheal plate have been observed, but whether they are homologous with the tracheids, or with the cells of the middle stony integument which they closely resemble, is uncertain.

One case was found of two nucelli within a common integument. A similar occurrence has been reported for Thuja (Coker, 1904), but in the present case it had definitely originated by the fusing of two integuments along their line of contact with a resultant failure to develop at this spot. The entire strobilus was distorted to such an extent that it was impossible to see whether or not the two ovules were located in the axil of a single sporophyll.

Another abnormality encountered was the development of more than one embryo-sac. Such a condition has long been recognised in Taxus and had been reported for Pinus, for example (G. Farmer, 1892). In two cases it was found in Phyllocladus. Oblique sections through a nucellus showing two megaspores in the free nuclear condition are shown in Fig. 54. These were from material of P. alpinus. A later case in P. glaucus was also obtained. Each embryo-sac was filled with endosperm and archegonia had developed in each, but these were disintegrating consequent on non-pollination. An interesting parallel was found in the case of a longitudinal section of an ovule of Pinus insignis. There were two female gametophytes

– 162 –

each with fully developed archegonia, fertilisation of those nearest the micropyle having taken place. The ovule was one of half-a-dozen which were being sectioned for class use.

While the general morphology of the ovule resembles the Taxineae more closely than it does the Podocarpineae these resemblances can be explained by placing Phyllocladus at the bottom of a line of evolution tending towards inversion of the ovule and consequent asymmetry of the arillus. Other important characters of the ovule link Phyllocladus more closely with the Podocarpineae. A case in point is the possession of a definite tapetal zone about the megaspore and the embryo sac. This zone is present in Phyllocladus, Podocarpus (Coker, 1907), Dacrydium (Young, 1907) and Saxegothea, but it is entirely absent among the Taxineae. Similarly, in its possession of a very definite two-layered megaspore membrane, Phyllocladus differs from the Taxineae which (Thomson, 1905) are peculiar in that they are the only tribe in which this membrane is a negligible feature. A megaspore membrane is also present among the Podocarpineae in Saxegothea (Stiles, 1908), Microcachrys and Dacrydium.

Endosperm formation in Phyllocladus is peculiar and is somewhat similar to what has been described for Ginkgo (Carothers, 1907). The cells are not laid down on the megaspore membrane, but have a free external wall of cellulose, although, as mentioned before, its exact homology with the endosperm membrane of Ginkgo is uncertain. Primary endosperm cells are formed and grow centripetally. Such a procedure has been described for Cryptomeria (Lawson, 1904), and in a modified way for Sequoia sempervirens (Arnoldi, 1899). Subsequent development in the three cases differs radically. In Phyllocladus periclinal wall formation follows each division of the nucleus of the primary endosperm cell without there being a comparable free nucleate stage with its subsequent independent cell-wall formation. Another difference is that while in Cryptomeria and Sequoia the primary endosperm cells are open towards the centre, in Phyllocladus they are closed. Endosperm development through the secondary formation of cells by centripetally growing primary ones is, however, common to all, and the development of the endosperm of Ginkgo as described by Miss Carothers (1907) can be interpreted in this way. It is also possible that this would explain the strongly radial rows of cells in the endosperm of Cycads (Chamberlain. 1906). In any case its occurrence in three such genera as Sequoia, Phyllocladus and Cryptomeria would justify further investigation of this point.

Two of the primary endosperm cells appear to act as archegonial initials, although secondary divisions in them may have occurred. This early differentiation of the archegonial initials is in accordance with the general trend throughout both Pinaceae and Taxaceae. In Phyllocladus and in all Taxineae the archegonial neck is never more than one layer of cells and is usually a rosette of four, although irregular divisions may result in a neck of from four to eight cells being formed. In Podocarpus, on the other hand, the archegonial neck may possess as many as three tiers of cells. The extreme

Picture icon

Fig. 1. Longitudinal section through a young strobilus of P. alpinus. The ovules are in the free nucleate condition of the megaspore. a—young arillus; a.s.—air spaces: m—micropyle; n—nucellus; t—beginnings of tracheal plate and basal palisade; v—vascular supply. × 30. Fig. 2. Median longitudinal section across the base of a young ovule as in Fig. 1. Basal plate of large nucleate cells which will differentiate as the basal palisade and tracheal plate. a—arillus; o—outer layer of integument; v—cells of the vascular supply. × 185. Fig. 3. P. alpinus. Longitudinal section through the micropyle of a young ovule. m—first cell of the middle stony layer to differentiate. Stained safranin and light green. × 250. Fig. 4. P. alpinus. Cells of stony middle layer in transverse section. See Fig. 8. Note characteristic crystals. × 250. Fig. 5. P. alpinus. Cells of stony middle layer in longitudinal section. o—cells of outer layer; i—cells of inner layer; c—crystals. See Fig. 7. × 250. Fig. 6. P. alpinus. Later stage in the development of the arillus. Appearance of hollowed structure. × 250. To face page 162

Picture icon

Fig. 7. P. Alpinus. Longitudinal section through an ovule when in a young embryo condition. a—arillus; e.m.—embroy; e—endosperm; g—gap in basal palisade; i.m. and o—the outer, middle and inner layers of the integument; n—nucellus; p—basal palisade; t—tracheal platform. × 30. Fig. 8. P. alpinus. Transverse section across an ovule as in Fig. 8. Lettering as above. × 30 Fig. 9. P. alpinus. L.S. showing the junction of the basal palisade and the middle stony layer of the integument. c—characteristic crystal; m—cells of the integument; p—cells of the pulisade: ti—tracheids of the tracheal plate. × 185. Fig. 10. P. alpinus. L.S. across the vascular gap in the basal palisade. Lettering as in 9 × 185. Fig. 11. P. alpinus. Tracheids of the tracheal plate showing characteristic taxinean sculpturing. Safranin and light green. × 325. Fig. 12. P. alpinus. T.S. across strobilar vascular bundle. ph—phloem elements; px—prot-xylem; xp—centripetal xylem; xf—centrifugal xylem. × 185.

Picture icon

Fig. 13. P. alpinus. Medial L.S. through young nucellus. ms—developing megaspore; mm—megaspore membrane; t—cells of tapetal zone. × 250. Figs. 15, 16, 17. P. alpinus. Developing megaspores in various stages of free nuclear division. That in Fig. 15 having a total of 4. although only 2 are shown in the section drawn. Lettering as above. × 250. Fig. 18. P. alpinus. Appearance of cells at apex of nucellus immediately prior to pollination. × 230. Fig. 20. P. glaucus. Median L.S. through the apex of a young endosperm showing two young archegonia. b—body cellor egg. nucleus; v—ventral canal cell nucleus; n—undivided neck cell. × 200.

Picture icon

Fig. 19. P. glaucus. Median L.S. through embryo sac during the formation of endosperm tissue. c—free external wall of cellulose; mm—megaspore membrane; p—primary radial cells; f—first periclinal walls. × 185. Fig. 21 P. glaucus. Young archegonium. e—egg nucleus; n—neck; mm—megaspore membrane; j—cells of jacket-layer. × 325. Fig. 22, 23, 24. P. glaucus. Unfertilised archegonia showing the development of the neck membrane. In Fig. 23 the cytoplasm around the egg simulates the appearance of that around a fusion-nucleus. Fig 22, 23 × 500; Fig. 24 × 325. Fig. 25. P. glaucus. T.S. through two unfertilised archegonia. Common jacket-layer and multinucleate condition of the jacket-cells. The ve nucleus shown actually occurred 70 μ higher up the archegonium. × 325. Fig. 26. P. glaucus. T.S. across the necks of the two archegonia in Fig. 25. Irregular division into a plate of six or seven cells. At h hautschicht formation appears to further subdivide the neck to give a plate of 8 cells.

Picture icon

Figs. 27, 28, 29. P. glaucus. Stages in the fertilisation of archegonia. Note the cytoplasmic sheath around the fusion nucleus. In Fig. 27 the 2nd. male nucleus is larger than usual, and in Fig. 28 the egg nucleus is not in median section and hence does not appear in its correct relative size. pt—pollen tube: n—neck cells: ma1 and ma2—male nuclei; es—cytoplasmic sheath; e—egg; fn—fusion nucleus. × 250. Fig. 30. P. alpinus. Mature archegonium. 250. Fig. 31. P. alpinus. Fertilisation. Lettering as in Fig. 27. vn—vegetative nuclei of pollen tube. X. 250. Fig. 32. P. glaucus. L.S. of apex of nucellus showing position and relative size of archegonia at time of fertilisation. m—micropyle; o—integument; pg—pollen grains; pt—pollen tubes; nu—nucellus; n—archegonial necks; mm—megaspore membrane; en—endosperm. × 70. Figs. 33, 31. P. alpinus, 8-nucleate state of proembryo. × 250.

Picture icon

Fig. 35. P. alpinus. L.S. apex of nucellus (nu) showing young embryo. Two vegetative nuclei (vn) may still be seen in the pollen tube (pt) and the second male nucleus (ma2) has not yet disintegrated. × 100. Figs. 36–41. P. alpinus. Young embryos in various stages of suspensor elongation. Apical cells may be traced in Figs. 36, 37, 39 and 40. × 250. Fig. 42. P. alpinus. Simple poly-embryony. One embryo shows very regular apical cell division. × 250. Figs. 43, 44. P. alpinus. Young embryos soon after the apical cell has ceased to function. × 250.

Picture icon

Fig. 45. P. alpinus. T.S. across embryo slightly younger than those in Figs. 43 and 44. × 250. Fig. 46. P. alpinus. Embryo which has just undergone a period of rapid sub-division. For explanation see text. × 185. Fig. 47. P. alpinus. Simple poly-embryony persisting to a late stage. × 185. Fig. 48. P. alpinus. Embryo showing the first anticlinal (a) and periclinal (p) divisions differentiating the cotyledons and the central cylinder. × 185. Fig. 49. P. alpinus. Embryo showing origins of cotyledons and root cap. a—stem apex; re—root cap. × 185.

Picture icon

Figs. 50, 51. P. alpinus. Embryo in resting condition. Fig. 51. is a median L.S. through the stem apex between the cotyledons. × 185. Figs. 52, 53. P. alpinus. Structure of the megaspore membrane of the mature ovule. Note many nuclei in cells of endosperm. o—oil droplets. Figs. 52 × 240, 53 × 325. Fig. 54. Oblique section through a young nucellus of P. alpinus in which two megaspores each in the free nucleate condition were developing. t—cells of tapetal zone; mm—megaspore membrane. × 250.

– 163 –

irregularity shown throughout the two families shows that this character cannot be taken into account phylogenetically. Phyllocladus seems to have advanced further than Podocarpus, but the condition of the neck in the latter is very probably influenced by physiological factors. The peculiar neck-membrane which had been described for P. alpinus is now seen to be present in P. glaucus also. This character seems to be purely physiological. In Phyllocladus, as in Taxus, Cephalotaxus and Podocarpus, a ventral-canal-nucleus is formed, but without cell-wall formation. In the character of the archegonial jacket Phyllocladus differs from the Taxineae, where it is much less conspicuous; but its condition among podoearps is not well enough known to institute comparisons. On summing up the characters of the archegonia it is clear that there is such variation in the two families and so many points of resemblance that on these grounds Phyllocladus could be assigned to neither.

The interpretation of the embryogeny of Phyllocladus is attempted along the lines laid down by Buchholz (1920).

In Phyllocladus alpinus wall-formation in the pro-embryo takes place after the division to give eight free nuclei. The pro-embryo consists of three tiers of cells situated at the bottom of the egg. The top tier elongates to form the suspensor, and one cell of the bottom tier functions as an apical cell. Eight free nuclei are also found in Podocarpus totara and Podocarpus nivale (Sinnott, 1913), sixteen in Cephalotaxus (Coker, 1907), Taxus (Jager, 1899) and Podocarpus coriaceous (Coker, 1902), but only four in Torreya (Coulter and Land, 1905). In the latter case the whole egg becomes filled with tissue. It will be noted that in Phyllocladus there is no trace of walled rosette cells, and obortive rossette nuclie seem also to be absent. The modifications of the embryo present in Podocarpus, for example, the cellulose cap in P. nivalis, and the terminal cap of cells in P. spicatus (Buchholz, 1920) are absent. Chamberlain in his review of Gymnosperms for the Botanical Review, June, 1935, mentions a paper to be published by Buchholz on the embryogeny of some dozen podocarps. I have been unable to see this paper. Chamberlain (loc. cit.) mentions that it establishes the occurrence of an embryo stage in most podocarps except Saxegothea and Microcachrys, in which the cells are binucleate. I have seen no trace of this in Phyllocladus, but it might be present in some of the very young embryos before the establishment of an apical cell. My material would not justify a statement on this point.

The embryo of Phyllocladus thus resembles most closely that of Taxus (Jager, loc. cit.), but no vestiges of cleavage polyembryony such as described for the latter have been observed, although simple polyembryony, persisting in some cases to a late stage, is the rule. Although, at present, it would be difficult to assign a phylogenetic position to Phyllocladus on the grounds of embryogeny, when more is known concerning the embryogeny of the various genera and even of the individual species, this should afford very valuable taxonomie evidence.

– 164 –

Summary of the Phylogenetic Affinities of Phyllocladus.

In the male sporophyll the winged pollen-grains and the male gametophyte Phyllocladus definitely shows podocarp affinities. In the character of the winged pollen-grains it is further advanced than Saxegothea and Microcachrys, but, considering the male prothallus, it stands nearer to these two genera and to Dacrydium than it does to Podocarpus. It also shows podocarp affinities in the possession of a definite megaspore membrane and a tapetal zone about the embryosac. The characters linking it with the Taxineae, namely, (1) Taxinean sculpturing of the tracheids, (2) centripetal xylem, (3) structure of the ovule and arillus, (4) resemblance of ovulate cone to that of Cephalotaxus are all superficial. The presence of Taxinean sculpturing and centripetal xylem are by themselves of small value, as their occurrence can be explained as being consequent on the extreme reduction of the genus. The structure of the ovule and arillus can, as has been seen, be correlated with a definite line of evolution in Podocarps, and the resemblances of the ovulate cone to that of Cephalotaxus must be more apparent than real both on account of the presence of two ovules in the axil of each sporophyll in Cephalotaxus and on account of the reduction experienced in Phyllocladus.

Summary.

I.

A detailed study of the general morphology of the ovule of P. alpinus is made and an account of the development of the integuments, arillus, basal palisade and tracheal plate is given.

II.

The development of the endosperm is described for the first time. (P. glaucus).

III.

The archegonia of P. alpinus are reinvestigated and a complete series of stages in the development of the archegonia of P. glaucus is described.

IV.

The embryogeny of P. alpinus is investigated.

V.

A general survey of the affinities of the genus is made and its phylogenetic position summarised.

VI.

In connection with laboratory technique a method for dealing with small ovules surrounded by hard integuments is developed and the effect of hydrofluoric acid on delicate structures examined.

Bibliography.
A. Laboratory Technique.

Chamberlain, C. J., 1932. Methods in Plant Histology, Univ. of Chicago Press, 5th Ed.

Couch, A. B., 1928. A method to soften tissue already embedded in paraffin, Science, 69, pp. 607–608.

Jeffery, E. C., 1917. The Anatomy of Woody Plants, pp. 449–451. Univ. of Chicago Press.

Land, W. J. G., 1915. Microtechnical Methods. Contributions from the Hull Botanical Laboratory, Bot. Gaz., 59, pp. 397–401

Langdon, La Dema M., 1920. Sectioning Hard Woody Tissues, Bot. Gaz., 70, pp. 82–84.

– 165 –

Rawlins, T. E., 1933. Phytopathological and Botanical Research Methods, p. 18, p. 25.

Tuan, H'su Chuan, 1930. Pieric acid as a destaining agent for iron-alum Haematoxylin, Stain Technology, 5, pp. 135–138.

Bibliography.
B. General.

Arnoldi, D., 1899. Beitrage zur Morphologie einiger Gymnospermen. I Die Entwickelung des Endosperms bei Sequoia sempervirens, Bull. Soc. Imp. Nat. Moscow, p. 13.

Baker and Smith, H. G., 1910. The Pines of Australia, New South Wales Govt. Press, pp. 413–431.

Buchholz, J. T., 1920. Embryo development and polyembrony in relation to the phylogeny of conifers, Amer. Journ. Bot., 7, pp. 125–145.

Carothers, Ida E., 1907. The development of the ovule and gametophyte in Gingko biloba, Bot. Gaz., 43, p. 116.

Chamberlain, C. J., 1906. The ovule and female gametophyte of Dioon, Bot. Gaz., 42, p. 321.

— 1935a. The Gymnosperms, Bot. Review, vol. 1, 6, pp. 183–208.

— 1935b. Gymnosperms, Structure and Evolution, Univ. of Chicago Press.

Cheeseman, T. F., 1925. Manual of the New Zealand Flora, pp. 119–121, 2nd Edition. N.Z. Govt. Press.

Coker, W. C., 1902. Notes on the gametophyte and embryo of Podocarpus, Bot. Gaz., 33, pp. 89.

— 1904. On the spores of certain Coniferae, Bot. Gaz., 38, p. 200.

— 1907. Fertilization and Embryogeny in Cephalotaxus Fortunei, Bot. Gaz., 43, pp. 1–10.

Coulter, J. M., and Land, W. J. G., 1905. Gametophytes and Embryo of Torreya taxifolia, Bot. Gaz., 39, pp. 161, 178.

Farmer, J. B., 1892. On the occurrence of two prothallia in the Embryo-sac of Pinus, Ann. Bot., 6, 213.

Jager, L., 1899. Beitrage zur Keuntniss der Endospermbildtung and zur Embryologie von Taxus baccata, Flora, 86, 241–288.

Jeffery, E. C., and Chrysler, M. A., 1907. The microgametophyte of the Podocarpineae, Amer. Nat., 41, p. 355.

Kirk, T., 1889. The Forest Flora of New Zealand, pp. 9, 195, 199, N.Z. Govt. Press.

— 1877. On the New Zealand species of Phyllocladus, Trans. N.Z. Inst., vol. X, pp. 378–383.

Kildaiil, M. J., 1908a. The Morphology of Phyllocladus alpina, Bot. Gaz., 46, 339–347.

— 1908b. Affinities of Phyllocladus, Bot, Gaz., 46, pp. 464–465.

Lawson, A. A., 1904. The gametophyte, fertilisation and embryo of Cryptomeria japonical, Ann. Bot., 18, p. 417.

— 1907. The gametophytes, fertilisation and embryo of Cephalotaxus drupacea, Ann. Bot., 21, pp. 1–23.

— 1910. The gametophytes and embryo of Sciadopitys verticillata, Ann. Bot., 24, 403.

Oliver, F. W., 1903. The ovules of the older Gymnosperms, Ann. Bot., 17, pp. 451–476.

Pilger, R., 1926. Taxaceae in Engler and Prantl's Die Naturlichen Pflanzenfamilien, 2nd. Ed.

Robertson, A., 1906. Some points in the morphology of Phyllocladus alpina, Ann. Bot., 20, pp. 259–265.

Sinnot, E. W., 1913. The morphology of the reproductive structures of the Podocarpineae, Ann. Bot., 27, pp. 39–82.

Stiles, W., 1908. The anatomy of Saxegothea conspicua, New Phyt., 7, pp. 209–222.

Thomson, R. B., 1905. The megaspore membrane of the Gymnosperms. Univ. Toronto Biol. Series, no. 4.

— 1908. Note on the pollen of Microcachrys, Bot. Geol., 46, p. 465.

Young, Mary S., 1907. The male gametophyte of Dacrydium, Bot. Gaz., 44, pp. 189–196.

— 1910. The Podocarpineae, Bot. Gaz., 50, p. 81.