Studies of New Zealand Nothofagus Species III—The Entire-Leaved Species
The two entire-leaved southern beech species of New Zealand, Nothofagus solanri and N. cliffortioides, although kept separate in botanical literature have always constituted a taxonomic puzzle. Many intermediate forms can be found in the field. Since the two species, as originally defined, have quite different ecological requirements, and their timbers have distinct properties, it is important that their taxonomic relationships be clarified. A study of leaf characters in populations throughout the range of the supposed species revealed a transition in most characters. The evidence was sufficiently conclusive to propose adopting the taxonomic procedure of making one species of the complete range, and retaining the originally described species as varieties.
In the first two papers of this series (Poole 1951, 1 and 2) detailed descriptions were given of the morphology of the flowers and fruit of the New Zealand species of Nothofagus. Earlier descriptions, including the original, were imperfect or inaccurate in these respects.
The species classification followed was that worked out by Cockayne (1926). This was based on leaf characters. Cockayne upheld with some reserve the two closely related, entire-leaved species, N. solandri and N. cliffortioides, remarking when so doing as to why “botanists have not considered them as two varieties of a compound species”. The morphological studies of flower and fruit provided no new characters by which these species could be satisfactorily differentiated, and in fact served only to substantiate their close relationship. In this paper, therefore, a re-examination has been made of the leaf characters by which these species were originally differentiated. These characters are most important, being those most frequently used in species identification.
The leaf characters that Cockayne used to distinguish N. solandri and N. cliffortioides and the minute differences in floral and fruit characters between them are set out below:
|N. solandri||N. cliffortioides|
|Leaf shape||Elliptic-oblong, cuneate base, obtuse, shortly petiolate.||Ovate - triangular irregularly rounded base, acute or subacute, shortly petiolate.|
|Size||1.2–1.5 cm long, 0.7–0.8 cm wide.||1.0–1.5 cm long, 0.7–1.0 cm wide.|
|Pistillate inflorescence||Very hairy; normally 2–, but sometimes 1–3-flowered||Viscous and almost glabrous, normally 2-, but frequently 1-flowered.|
An examination of trees in the field discloses that only some fit readily into the above descriptions. As far as leaf characters are concerned they constitute a puzzling pair of species Although kept separate in floras, many intermediate leaf forms can be found. For these reasons confused identifications have often been recorded in literature and in herbaria. It is important that the relationship of the two species be worked out because they have very diffeernt ecological requirements.
The only method of sorting out a problem of this nature is by studies of populations over as wide a range as possible of the plants' distribution. It is the interbreeding population that goes to make up the species and it is the population that must be studied.
This paper gives the results of the analysis of twelve populations occurring throughout the combined ranges of the two species. It should be noted that, throughout the description of this analysis, the two species names are used in the same sense that Cockayne used them. The final result of the analysis, however, shows that there is a genetic continuum, and the two species names, as used hitherto, represent the two extremes In the concluding part of this paper therefore the whole continuum is placed under one species and the two extremes are retained as varieties.
General Remarks on Variation of Leaves
The leaves of seedlings and saplings are quite different from those of larger trees, but, because they vary both from plant to plant and with the age of the plant, the study of them is highly complicated and has not been included here. Such a study would, almost certainly, involve growing seedlings and saplings.
The leaves of a mature tree vary in shape and texture according to their position and exposure on the tree or their position on the twigs that bear them; and they vary in size from season to season or between spring and late summer shoots Thus leaves from the upper part of the crown, where there is the maximum degree of exposure, are smaller and more coriaceous than those from the lower part; and those from the inside of the crown are larger and thinner than those from the outside at the same level. Shade leaves are glabrous instead of tomentose on the under-surface, and reversion shoots carry leaves having seedling or sapling characters according to the height from which they arise on the trunk Finally, on any one twig, leaf shapes differ, especially the shape of the terminal leaf, which is usually broader than the others.
The general leaf shape and size also vary considerably from tree to tree in any stand.
Method of Study
(a) Leaf Characters
The separation of the two species by leaf characters has hitherto been based on the general shape, the nature of the base and tip and the average size of the leaves. In this study all the above characters have been analysed; in addition two characters were chosen for statistical analyses, the ratio of the lamina-length to breadth, and the leaf-base angle (see explanation below). These two characters were the only measurable ones and seemed to be important when analysing general shape. Two other characters were examined in some detail, the hairiness of the petiole and longitudinal ridges which are sometimes present between the midrib and sides of the leaf.
Leaf shapes were recorded under six categories; roundish, ovate-triangular, elliptic-ovate, elliptic, elliptic-lanceolate and elliptic-oblong. Different observers would classify the same leaves under different shapes to a limited extent, but it is considered that the same observer would be uniform throughout his classification so that the trend in differences would be accurately recorded.
Leaf-bases were recorded as being even when the lamina joined the petiole at the same level on either side of it, and uneven when the lamina joined the petiole at different levels on each side.
Leaf apices were described as being acute or obtuse, there being a fairly clear-cut division between these two types though some workers might have seen fit to use the category sub-acute as well.
Size of leaves has always been given in literature as the length and the breadth of the lamina. These measurements were used as indicating size in this study.
Ratios were calculated from the measurement of the greatest lengths of the laminas divided by the greatest widths. They were figures that could be used statistically to show the difference between leaf shapes; an increasing ratio represented leaves becom-
ing relatively narrower and indicated the transition sequence, roundish, ovate-triangular, elliptic-ovate, elliptic and elliptic-lanceolate or elliptic-oblong.
Leaf-base angles were the angles between tangents to the leaf margins where they joined the petioles. The measurements expressed precisely the difference between the “cuneate base” of N. solandri and the “rounded base” of N. cliffortioides which in turn had a bearing on the “elliptic-oblong” shape of the former and the “ovate-triangular” shape of the latter.
(b) Sampling of Individual Trees
A sampling method was derived from a study of the above characters of leaves from different parts of an individual tree. This consisted of taking ten twigs from the lower outside part of an open-grown or fairly open-grown tree. From each twig the leaf next to the terminal one was taken for examination. Thus ten leaves were taken from each tree for study and analysis of the characters listed above.
(c) Sampling of a Stand
Significant differences, between some leaf characters, from tree to tree were shown by analysing samples, collected as above, from eleven trees in a stand of N. cliffortioides in the “Ten Mile Bush”, Te Anau. In shape, all the leaves were ovate-triangular with uneven bases and acute tips. An analysis of variance of the ratios and basal-angles on the other hand, showed that trees differed significantly in their characters.
To the eye, distinct differences between the leaves of one tree and another were discernible, and in this particular stand these differences could be recorded by measuring the basal angle, the length and breadth, and calculating the ratio, for a sample of leaves from each tree. In other stands it was possible to record differences between trees by the differences in leaf shape, base and apex as well.
A stand therefore had to be sampled by collecting leaf samples from a number of individual trees. Between forty and fifty was the number chosen, as erring on the safe side, and being a physical practicability. A smaller number might have been sampled, but this was not known until the whole study was completed.
(d) Sampling throughout the Range of the Two Species
To determine differences between stands, collections were then made from forests in twelve areas, forty to fifty trees in each stand being sampled to cover the range in variation in any one stand. The twelve areas covered a wide geographical range, the need for which is explained in the following notes on the distributions, hitherto accepted, for the two species.
Cockayne defined the distribution of N. solandri as follows: a northerly limit on the Mamaku Plateau; then common in the Ruahine-Tararua-Rimutaka chain always in the lowland and montane belts, on the old coastal plain in the east of the Egmont-Wanganui district, and in the Sounds-Nelson district, Banks Peninsula and the Oxford and Alford forests of Canterbury. “More than the above, I cannot say, except that it is certain that, again and again, N. cliffortioides has been mistaken for N. solandri and published in error.” He defined the distribution of N. cliffortioides as essentially high-mountain, with its northerly limit in the East Cape district, thence following the
main mountain chains of the North and South Islands except for the Tararua Mountains. In the southern half of the South Island, it descended to sea-level.
Accepting, meantime, the above distribution, it will be seen that the two species overlap in latitude from about the middle of the North Island to Alford on the east of the South Island. As far as altitude is concerned, where one occurs at high elevations the other is often present at lower elevations on the same mountains.
(e) The Forests Sampled
The following are brief descriptions of the forests from which the twelve collections were made. The specific names used are those by which the whole stands were identified at the time of collection. The locations of the forests are shown in Fig. 1.
Butterfly Creek. Eastern Port Nicholson (Harbour to Wellington). The floor of a wide valley, occupied in places by stands of pure N. solandri interspersed with trees or stands of N. truncata. This forest has been partly burned, and the edge now consists of a pure stand of second-growth N. solandri. The collection was made from trees within the pure N. solandri.
Picton. Remnant stand of mixed N. solandri and N. menziesii.
Northern Tararuas. Remnant stand of N. solandri.
Alford Forest. Almost pure entire-leaved beech forest.
Pokaka. Mixed entire-leaved beech, broad-leaved, podocarp forest.
Te Rei Forest, Southern Kaimanawa Mountains. This forest, on the edge of the Central North Island pumice showers, is composed principally of Nothofagus fusca at higher elevations mixed with podocarps and broad-leaved species at lower elevations. The whole forest, where undisturbed, has an edge (ecotone) abutting on tussock grassland, of entire-leaved beech. Collections were made at different elevations in this edge.
Te Anau, Ten Mile Bush. Forest of N. cliffortioides.
Thomas River, Canterbury. Forest of N. cliffortioides.
Central Kaimanawa Mountains. Forest of N. cliffortioides.
Presentation of Data
Table I gives all information and data for the twelve collections. Leaf shapes have been recorded as the percentage distribution of each shape in a collection, and the base and apex as the percentage distribution of the different types. The mean, range, and standard deviation for leaf ratio and basal angles have been given for each collection.
Since the twelve collections were gathered from different latitudes and altitudes, a clearer idea of the relative variations of characters was obtained by using Zotov's (1938) climatic belts and relating to them the collections made throughout New Zealand. The belts with which we are concerned here he named, in ascending altitude, “lower and upper warm temperate” (0–600 metres), “lower and upper cold temperate” (600–1,200 metres). These belts descend in altitude from the north to the south of the country as shown in Fig. 2 If the altitudes at which collections were made were related to a common latitude, it was possible to compare more nearly one collection with another. For example the Thomas River, Canterbury, collection at latitude 43.5° S. and altitude 820 metres would be projected to 450 metres at 40.5° S., and so on. In Figs. 3 and 4, using this method, the basal angles and leaf ratios have been plotted against altitudes at the common latitude of 40.5° S.
Fig. 5 illustrates a selection of leaves, each representing an individual tree, showing the variation within and between populations.
Analysis of Data
(a) Leaf Shape
The populations 9, 11 and 12, representing N. cliffortioides, had predominantly ovate-triangular leaves: none of these populations had elliptic-oblong leaves. The populations, 1, 2 and 3, on the other hand, had predominantly elliptic leaves with fair proportions of elliptic-ovate or elliptic-oblong; but none had ovate-triangular leaves. The unclassified populations, 4, 5, 6, 7 and 8 all had predominantly elliptic-ovate leaves, as did the N. cliffortioides population from Te Anau. All, except No. 7, had some ovate-triangular leaves, and four of the populations had elliptic-oblong leaves.
There was, therefore, a shift in emphasis from elliptic shapes in N. solandri to elliptic-ovate in the unclassified populations and ovate-triangular shapes in N. cliffortioides. While the leaves in the populations of the two species were clearly distinct, they were connected by a series of intermediates in the unclassified populations which also contained leaves of the two species.
This composite measurement, which was closely related to shape, was greatest on the average for the N. solandri populations and was least for N. cliffortioides populations. The unclassified populations had various intermediate mean ratios. Considering the range within each population, however, there was a considerable overlap between all populations (Fig. 4). For example the range in No. 1 was from 1.44 to 2.83, while that in No. 9 was 1.44 to 2.04. The lowest ratio was associated in No. 1 population with roundish leaves, while in No. 9 population it was associated with ovate-triangular leaves.
(c) Basal Angle
This character was also closely related to shape, since it was a measure of the “cuneate” leaf base of N. solandri and the “rounded” leaf base of N. cliffortioides.
The means for the three N. solandri populations ranged about 80°, and for the N. cliffortioides populations from 133° to 138°, excluding No. 10 (Te Anau), which had a mean of 112°. Unclassified populations had intermediate means. The marked degree of overlapping in the range of the measurements between populations is shown in Fig. 3.
(d) Leaf Base
In the N. solandri collections the largest proportion of the leaves had even bases, but the N. cliffortioides leaves nearly all had uneven bases. Unclassified populations were variable in this respect.
(e) Leaf Apex
Leaves of the N. cliffortioides populations were mostly acute, whereas those of the N. solandri and unclassified populations were mostly obtuse.
(f) Leaf Size
Excluding the Te Anau population, there was a diminution in the mean lamina length from N. solandri to N. cliffortioides populations, but little difference in lamina width. The leaves of the Te Anau population had the largest mean lamina length and width.
(g) Other Characters
The longitudinal ridge between the lamina margin and midrib was frequently present in the leaves of N. cliffortioides populations; it was never present in N. solandri populations. They were occasionally found in the leaves of unnamed populations, but then only when the leaves had other N. cliffortioides characters.
L, length; B, breadth; R, ratio (L/B); BA, basal angle.
(h) Correlation of Characters
In Cockayne's descriptions of the species the general leaf shape is correlated with the nature of the leaf base and leaf apex and with the average size. This study has shown that these correlations did exist for the majority of trees in any population of the two species but not for all of them. In the unclassified populations, this same group of characters was often correlated, but the leaves of many trees had different combinations and also had many characters intermediate between those of the two species.
The statistics, length, breadth, leaf ratio and basal angle of ten populations were analysed for correlation coefficients. These are given in Table II.
The character of uneven leaf-base was more commonly associated with other N. cliffortioides characters, and the lateral ridge was invariably correlated with other N. cliffortioides characters.
(i) Variation with Altitude
Fig. 5 shows clearly the gradual change, with considerable over-lapping in each succeeding population, from the relatively lower to the higher altitudes. The elliptic-oblong, elliptic-ovate and elliptic leaves of the N. solandri populations from the lowest elevations give way to the ovate-triangular leaves of N. cliffortioides at the highest elevations. The elliptic type leaves in populations 1 and 2 are to be found in the populations 5, 7 and 8, from intermediate elevations, which also contain ovate-triangular leaves. The peculiar roundish leaves in 1 and 2 are also found in 5 and 8.
The gradual change in the means of the ratios and basal angles accompanying changes in elevations of the populations is shown in Figs. 3 and 4. There is a great deal of overlap, from population to population, in the amplitude of these measurements.
Before finally discussing the relationship of the two species on the basis of the evidence presented here, it is necessary to consider the place that inter-specific hybridisation and introgression could play in affecting intra-population variability.
It is well established that inter-specific hybridisation takes place between the pairs of species N. cliffortioides and N. fusca, N. solandri and N. truncata. Fl trees, with intermediate leaf forms can readily be located in forests containing these two pairs of species, and it is known that F1 trees are fertile and produce segregating progeny (Poole 1951). Cockayne also stated that hybrids occurred between combinations of any two of these species. Field observations would support this, and an artificial hybrid has been produced between N. solandri and N. fusca (Poole 1951).
This type of hybridisation, if between greatly differing species, gives distinct F1's and hybrid swarms. Nevertheless the bulk of a forest containing such a pair of species is composed of the species themselves; there is no suggestion of a transition of any character from one species to the other as has been shown in this study between N. solandri and N. cliffortioides. One species could, however, affect the other by introgression, but to what extent this happens has never been studied critically.
Incipient toothing of entire-leaved species is probably the most obvious evidence of hybridisation. Amongst the samples collected, close examination showed that some trees had leaves with this character. For example collection No. 9, Te Rei, Fig. 5. Hybridisation between N. cliffortioides and N. fusca was present in this forest, and back crossing could have produced the irregular margin seen in the illustration of the fifth leaf from the left. On the other hand, very fine toothing is to be seen in the illustrations of collections 10 and 12 which represent collections from pure N. cliffortioides forest remote from any N. fusca. It would be possible for N. fusca genes to have entered these populations from trees that have disappeared with changing climate.
If toothing is a character that can get into a population by introgression, other characters must do so too. On the evidence available at present, it seems reasonable
Fig. 4.—Ratios of leaf collections plotted against the common altitude of 40.5° S. The range, mean and standard deviation of each collection is shown.
to assume that introgression takes place, though to what extent it would be impossible to say without much detailed study. On the other hand, as great an order of variation is to be found in a population such as No. 4, Alford, which is remote from other species.
It seems that we are dealing with geographical or clinal variation such as occurs in many plants, particularly forest trees.
Notes on Each Population
The following are brief notes on each collection, including remarks on individual leaf characters, correlations and transitions of both single characters and combinations of them.
No. 1—Wellington. A population most typical of Cockayne's N. solandri. Elliptic shapes commonest (second highest mean ratio); leaf-bases cuneate (lowest-mean basal angle). Transition of shapes from elliptic-lanceolate to rotund, which shape could be the result of N. truncata influence. No sample could be mistaken for N. cliffortioides.
No. 2—Picton. Somewhat similar to No. 1, but with shapes ranging through to ovate-triangular, which leaves could be identified as N. cliffortioides. In this connection, N. cliffortioides comes to low elevations in the Nelson-Marlborough district.
No. 3—Tararuas. Similar to the Wellington population in shapes and measurements, and as with that population the only possibility of hybridisation could be with N. truncata.
No. 4—Alford. Compared with the three previous collections this population has a smaller mean ratio, smaller leaves and a greater mean basal angle. Some leaves would be identified as N. solandri and some as N. cliffortioides.
No. 5—Pokaka. Shapes vary from elliptic-oblong to ovate-triangular, there being a fair percentage of the latter. Leaves generally smaller, mean ratio smaller and mean basal angle greater than in any of the previous collections. Many leaves would be identified as N. cliffortioides but a few as N. solandri. The entire-leaved beech forest at Pokaka is, or has, before the destruction of much forest, been continuous with N. cliffortioides forest on the flanks of Mt. Ruapehu.
No. 6—Te Rei A. 600 metres.
No. 7—Te Rei B. 600 metres.
No. 8—Te Rei. 800 metres.
No. 9—Te Rei. 1,000 metres.
Numbers 6, 7 and 8 are comparable in shapes but show decreasing sizes and mean ratios. Although all three are at equal or greater elevations than the Pokaka population, less N. cliffortioides is apparent. In all of them leaves could be identified as either species, but the majority would be unnamed. The three populations could be influenced by N. fusca.
Number 9 was identified as N. cliffortioides. The mean size of the leaves was the same as that of Number 8, the mean ratio was smaller and the basal angle much higher. There was ample evidence that this population could also have been affected by N. fusca.
No. 10—Te Anau. This population, identified as N. cliffortioides has a number of exceptional features. Trees are of good form, with tall, straight boles, characters usually associated with N. solandri. Nearly all leaves, however, would be identified as N. cliffortioides, but, on the average, they are the largest leaves of any population. The leaf ratio is that of other N. cliffortioides collections, but the basal angle is intermediate. Leaf apices are mostly obtuse, in contrast to N. cliffortioides leaves generally. This population is typical of what is called lowland Southland N. cliffortioides. As stated earlier, if incipient toothing can be accepted as evidence, there is some influence of N. fusca in the population. It might be a possibility that this good-formed N. cliffortioides is the result of introgression by that species. It should be noted that the population is continuous with more typical N. cliffortioides at higher altitudes in the same region.
No. 11—Thomas River, Canterbury.
No. 12—Kaimanawa Mts.
These two collections, although very widely separated, are similar in most details and all leaves would be identified as N. cliffortioides.
Validity of the Species
Because of the transitions of individual leaf characters and of combinations of characters, it is only possible to apply, with certainty, the present names at either end of the variability scale; anything between the extremes cannot be named. Some forests contain trees with mainly these intermediate leaf forms, such as, for instance, the forest from which collection No. 5 was obtained. In this category also fall many of the forests mentioned by Cockayne as containing N. solandri; the forests of the Mamaku Plateau, Egmont-Wanganui and a great part of the Sounds-Nelson, north-west South Island and Canterbury foothills at lower and intermediate elevations. All the entire-leaved beech south of Alford has been called N. cliffortioides, but many leaves in the Te Anau collection are intermediate.
It could, of course, be considered that characters, other than those concerned with leaves, could be chosen for identification purposes, but, at present, it seems unlikely that suitable ones will be found. Identification must continue, for the time being then, to be based primarily on leaf form.
In spite of continuous variation, it seems that the identity of the two species, N. solandri and N. cliffortioides, should be retained by some means because the extremes which these represent have greatly different ecological requirements. But,
by using the present nomenclature we are faced with the impossibility of naming trees with intermediate leaf-forms. The most satisfactory solution to this difficulty seems to be to use the method of applying one specific name to the whole range of leaf-forms, giving varietal status to the two present species.
This proposal is put forward with some misgiving, because, to abide by the rules of botanical nomenclature, it is necessary to use one of the present names as the new specific name. For the botanist, to whom the rules are part of his science, this presents no difficulties; but the utmost confusion will reign for the forester and sawmiller, who are even more concerned than the botanist in the accurate naming of these beeches.
Renaming the Species
Nothofagus solandri (Hook. f.) Oerst (emend nov.)
A tree, dwarfed to a much branched shrub at high elevations, growing to 27 metres high and to one metre or more in diameter at lower elevations. Bark pale and smooth on smaller or young trees to black and roughly furrowed on larger trees. Leaves evergreen, entire, coriaceous, venation obscure, shortly petiolate, lamina 1–1.5 mm long, elliptic-oblong to ovate-triangular, apiculate or rounded at the tip, greyish-white or fulvous tomentum beneath. Flowers monoecious; both male and female borne on new spring shoots, the male at the bottom of the shoot and the female inflorescences at the top and axillary in the new leaves; male sessile, with 2–3 mm wide, bell-shaped, membranous perianth containing 8–17 dark-red stamens; female inflorescence sessile, 1–2 mm wide, perianth hairy or viscous and glabrous, containing 1–3 minute di- or tri-merous flowers. Cupules 3–5 mm long, with 3 segments, one usually shorter than the others, each segment with two rows of bracts. Nuts oval in outline, flat or triquetrous, 3.5–6 mm long, dark straw.
Occurring from about latitude 38° S. to Foveaux Strait; sea-level to timber-line.
var. solandri Black beech. A tree from 14–27 metres high and .5–1.5 metres in diameter Bark of young trees pale and smooth, of older trees black and roughly furrowed. Leaves elliptic-oblong to orbicular, mostly cuneate at the base, rounded tip. Female inflorescence very hairy. Occurring in the East Cape, Egmont-Wanganui and Ruahine-Cook Botanical Districts of the North Island, and the Sounds-Nelson, North Eastern and Eastern districts of the South Island, at elevations only up to 600 metres, but usually much lower.
var. cliffortioides (Hook. f.) Poole comb. nov. F. cliffortioides Hook. f. in Hooker Icones Plantarum T. DCLXXIII.
Mountain beech. A tree up to 20 metres high, but reduced to a shrub at high elevations, and with a trunk up to 70 cms in diameter. Bark pale and smooth Branchlets often distichous. Leaves ovate-triangular to elliptic-ovate; bases wide-angled and mostly uneven; usually acute, sometimes apiculate; many trees having some leaves with a peculiar lateral ridge on the upper surface between the margin and midrib. Female inflorescence viscid. In the North Island, occurring at high elevations from Lat. 38° S. to the northern Ruahine Range and throughout the South Island from low elevations up to timber-line, but commonest at high elevations.
A large-leaved form, growing as a well-formed forest tree up to 25 metres high, is present in fair quantity in lowland western Southland.
One does not lightly propose a change in the names of common forest trees. For this reason the examination of field material and preparation of the paper have been spread over an inordinately long period, during which many people have assisted with leaf measurements and have critically read the manuscript at different stages.
Thanks are due to all these and especially to L. B. Moore, E. J. Godley and J. T. Holloway for criticisms and suggestions, and to M. H. Bannister for criticisms, assisting with leaf measurements and the drawing of Plate 5.
Cockayne, L., 1926. Monograph of New Zealand beech forests. Part I. The Ecology of the Forests and Taxonomy of the Beeches. N.Z. State For. Bull., 4. Govt. Printer, Wellington.
Poole, A. L., 1950. Studies of New Zealand Nothofagus Species. 1. Taxonomy and Floral Morphology. Trans. Roy. Soc. N.Z. 78, 363–380. 2. Nut and Cupule Development. Trans. Roy. Soc. N.Z. 78, 502–508.
—— A. L., 1951. Hybrid Southern Beeches. N.Z. J. of For. 6, 3.
Zotov, V. D., 1938. Some Correlations between Vegetation and Climate in New Zealand. N.Z. J. Sci. Tech., 19, 474–87.
A. L. Poole,
State Forest Service,
Fitzherbert Terrace, Wellington, N.1.