Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 85, 1957-58
This text is also available in PDF
(1 MB) Opens in new window
– 227 –

Variation in Samples of Two-year-old Pinus attenuata, P. radiata and their Hybrids

[Received by Editor, August 21, 1956.]


Samples of two-year-old Pinus attenuata and P. radiata displayed many morphological differences, particularly in stature, branching, development of buds, foliage, and colour. Smaller, but still distinct, differences were noted between three samples of P. attenuata from widely separate places in California.

F1 hybrids between the species were intermediate in most characters. Seed collected from a single putative hybrid in Ashley Forest, Canterbury, gave rise to a strongly segregated hybrid swarm.

Analysis by hybrid indices, using seven characters, effectively discriminated the two species, placed the F1 hybrids half-way between the species, and suggested that the hybrid swarm was mainly the result of back-crossing to P. radiata.


In 1953 some small samples of seed were received from the Institute of Forest Genetics in California. Among them were collections from the following sources:–

  • (1) Pinus attenuata Lemmon. More than one tree at Ensenada, Baja California, Mexico.

  • (2) P. attenuata. A single tree at 4,370ft elevation, on the south slope of Mt. Shasta, about two miles NW of McCloud, California.

  • (3) P. attenuata. A single tree referred to as Eld-2=1, a few miles from the Institute of Forest Genetics, Placerville, California.

  • (4) Pinus × attenuradiata Stockwell and Righter. The result of artificial pollination of the one P. attenuata parent (Eld-2=1,) with pollen from two P. radiata trees in the arboretum at Placerville.

During his visit to New Zealand in 1952, Dr. J. W. Duffield had noted in Ashley Forest, Canterbury, a tree which he regarded as a hybrid between P. attenuata and P. radiata, and seed from this tree was also received.

All five lots were sown in the spring of 1953. The seedlings did not attract attention at first, but in the winter of 1954 a striking contrast was noted between the P. attenuata from McCloud and those from Placerville. The foliage of the former was tinged a dull purple, the stems were all bent to one side near the ground, the tuft of apical leaves on each was inclined instead of erect, and the tips of these leaves were curved in the same direction; the whole appearance was lax. The P. attenuata from Placerville had erect stems, the foliage was greyish and only slightly coloured with purple, and the leaves were stiff and straight.

In the winter of 1955, the five lots of seedlings displayed such impressive differences that a morphological analysis seemed practicable. A sixth sample was added from adjacent P. radiata seedlings which had been grown from a Rotorua seed-lot under the same conditions.

The essential similarities and differences between the six samples are summarized in Table I, and representative individuals are illustrated in Plate 19. So distinct were the first five samples (i.e., all except the one from Ashley) that a competent observer, who was familiar with the differences between them, could have taken a thorough mixture of individuals and sorted them correctly into their five original groups.

– 228 –

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Table I.—Showing General Resemblances and Differences Between Young Trees from Six Different Seed-Lots.
Taxonomic Status P. attenuata Interspecific F1 Hybrids P. radiata Hybrid Swarm
Place of Origin. Ensenada McCloud Placerville Placerville Rotorua Ashley
Stem Tall, rather crooked Short, rather crooked Short, rather crooked Very tall, almost straight Very tall, straight Very variable
Branches (long shoots) Medium-long Long Long Medium-long Short Very variable
Young shoots Yellow-brown to greenish, scarcely shining Dark grey-brown, shining Yellow-brown, shining Greenish brown, rather shining Green, not shining Very variable
Density of dwarf-shoots Sparse Medium Sparse Dense Dense Very variable
Needle-leaves Pale, dull yellow-green Very lax Very long, thick Bluish, grey-green Stiff Long, very thick Dull grey-green Stiff Long, thick Green Lax Very long, thick Dark green Rather lax Short, thin Dark green to pale yellow-green Lax, some very Dimensions very variable
Scale-leaves Brown Brown, dark at edges Brown, very dark at edges Brown, dark at edges Uniformly brown Very variable
Purple colouring in winter Weak Very strong Strong Weak Absent Very variable
Stem apices Leafy, or in imperfect buds Mainly in buds All in perfect buds; narrow, resinous Mainly in buds; fat, rather soft Leafy Very variable
– 229 –

Morphological Analysis of Single Characters

Random samples of the six populations were taken during August. Fifty trees of P. radiata were lifted and recorded in the laboratory, taking care that they did not become dried and shrivelled. The other samples consisted of 10 trees each for the three P. attenuata lots, 10 for the F1 hybrids, and 50 for the progeny from Ashley Forest. These trees were recorded in the nursery or in the laboratory just before they were planted out. The characters recorded were:

  • (1) Stem Height

    This was measured from the level of the cotyledons to the tip of the bud or apical tuft of primary leaves. No attempt was made at statistical comparisons, but there were some obvious differences, as indicated in Table 1.

  • (2) Number of Pseudo-whorls of Long Shoots on Stem

    This was usually four in all lots except the Mexican P. attenuata, most of which had three.

  • (3) Number of Shoots in each Pseudo-whorl

    P. attenuata had, in general, rather fewer shoots in each pseudo-whorl than P. radiata, but the distinction was slight.

  • (4) Length of Lateral Shoots in Relation to Stem Height

    By plotting the length of the longest lateral shoot against stem height for each tree, it was evident that the two species differed P. attenuata had relatively longer branches, although the Mexican lot differed little from P. radiata. The ratio Length of longest branch/Height of stem was therefore used as a branching index for each tree (Text-fig 1).

  • (5) Proportion of Stem Bearing Dwarf-shoots

    The stem can be divided into alternating parts distinguished as those which bear dwarf-shoots and those which do not. By expressing the total length of the sections bearing dwarf-shoots as a percentage of stem height, a good contrast was found between the P. attenuata from Placerville and P. radiata; the F1 hybrids were intermediate but resembled P. radiata the more closely of the two. When the other lots of P. attenuata were examined, however, such a distinction was no longer clear.

    Picture icon

    Text-fig. 1.—Frequency distributions of four characters in samples of six different populations. Further details in text.

– 230 –
  • (6) Density of Dwarf-shoots on Stem

    The number of dwarf-shoots on 10 cm of stem was counted for each tree (except 15 of the P. radiata sample). The count was made about mid-way between base and apex, and any sections devoid of dwarf-shoots were excluded. In a few cases only 7.5 cm of stem was suitable, and the count was scaled up accordingly. In P. radiata nearly every node of a series had an axillary dwarf-shoot, but in P. attenuata there were many such nodes without visible axillary structures. It was this, rather than a difference in the length of the internodes, which resulted in the general contrast between the dense foliage of one species and the sparse foliage of the other, (Text-fig. 1.)

  • (7) Primary Leaves

    At first, the mean length and mean breadth of 10 leaves per tree were calculated. The Mexican P. attenuata appeared to have longer and broader leaves than any of the other samples, but the dimensions were excessively variable and failed to distinguish the species.

  • (8) Scale-leaves

    These, the scarious homologues of the primary leaves, were usually associated with apical buds, but were also found in places further back on the stem. They were not suitable for measurement, and were not found on every tree, but they did show a sharp contrast in colour between the Placerville P. attenuata and the P. radiata. In the former, as well as in some mature specimens of P. attenuata in New Zealand, they were dark brown, especially at the margin and apex, which were sometimes almost black. In P. radiata they were paler brown and more uniformly coloured. This specific difference seemed not so strong for the McCloud sample and was even weaker for the Mexican one. The progeny from Ashley Forest showed a complete range between the extremes; there were more individuals like P. radiata than those like P. attenuata, but most were intermediate.

  • (9) Number of Needle-leaves in Ten Fascicles

    For this and all the other characters of the needles, samples of 10 fascicles each were removed from the stem of each tree, taking reasonable care that each sample was representative of the cauline foliage as a whole. Samples from some trees included one or more immature fascicles; other samples contained only mature fascicles. Although it might have been better to restrict the sampling to, say, the middle third of the stem, the method proved satisfactory.

    P. radiata often had 3 needles to a fascicle, very often 4, and sometimes 5 or more; P. attenuata had usually 3, and sometimes 4; the F1 hybrids were intermediate; the Ashley hybrids were also intermediate but more like P. attenuata.

  • (10) Mean Length of Ten Needles

    The ranges and means (in cm) for this character in the six samples were:–

    P. radiata 7.8 – 10.2 – 12.2
    P. attenuata (Placerville) 9.3 – 10.9 – 12.4
    P. attenuata (McCloud) 9.1 – 11.0 – 12.6
    P. attenuata (Ensenada) 11.0 – 12.5 – 13.4
    P. × attenuradiata (F1) 11.2 – 12.4 – 14 0
    Hybrids (Ashley Forest) 8.8 – 11.1 – 13.9

    This indicates that the P. attenuata, especially the Ensenada lot, had longer needles than P. radiata. The F1 hybrids had very long needles; this was perhaps a symptom of heterosis.

  • (11) Mean Breadth of Ten Needles

    The greatest breadth of one needle in each fascicle was measured to the nearest 0.1 mm. This character was somewhat correlated with needle length, as one might expect, but it was the more efficient of the two for discriminating the species. (Text-fig. 1.)

  • (12) Spacing of Teeth on Needle Margins

    The central part of one needle of each fascicle was laid on the stage of a binocular microscope at low magnification, with one margin occupying a diameter of the field (1.12 cm). In this way counts of marginal teeth were made, and the mean calculated, of 10 needles from each tree. P. attenuata had more widely spaced teeth than P. radiata (Text-fig. 1). Scatter diagrams of tooth counts plotted against needle length suggested that within each species the two characters were quite independent of one another.

  • (13) Colour and Appearance of Young Shoots

    For this rather composite character, individuals were scored in subjective grades according to their position in a scale ranging from “pale green, not shining” (typical P. radiata) to “yellow-brown, shining” (extreme P. attenuata). The contrast between P. radiata and the P. attenuata from McCloud and Placerville was strong, and the F1 hybrids were intermediate. The Mexican P. attenuata, however, was intermediate or like P. radiata. The hybrids from Ashley Forest were mostly intermediate, but varied more towards “green, not shining” than towards the other extreme.

  • (14) Winter Colouring of Foliage

    This was an outstanding character, because the foliage of P. radiata was entirely and invariably green, whereas in every other lot there was some degree of purple tingeing. This colour appeared in autumn and disappeared in spring. It was strongest distally and usually faded out proximally in relation to individual leaves and to the shoots. From a distance, in combina-

  • Picture icon

    Two-year-old individuals showing some of the differences between six populations. The background is ruled in 5 cm squares. Figs 1–3—Pinus attenuata (1, from Placerville, 2, from McCloud. 3, from Ensenada) Fig. 4—An F1 hybrid, P. attenuata × P. radiata Fig. 5—P. radiata from Rotorua Figs 6–9—Four hybrids, showing extreme segregation towards P. radiata (6, 7) and P. attenuata (8, 9)

– 231 –
  • [The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

    Table II.—Showing Method of Allotting Values to Seven Characters for Calculating Hybrid Indices.
    Character. Index Value.
    0 1 2 3
    Purple colouring in winter Absent Faint to moderate Strong, conspicuous
    Development of bud on stem apex None Imperfect, scales loosely imbricated Perfect
    Mean needle breadth (mm) 0.6–1.1 1.2 1.3 1.4–1.8
    Ratio of length longest branch: stem height 0.05–0.27 0.28–0.37 0.38–0.75
    No. of dwarf-shoots on 10 cm of stem 41–80 31–40 10–30
    Mean number of teeth on 1.12 cm of needle margin 35–45 32–34 25–31
    Appearance of young shoots Pale green, not shining Intermediate Brown, shining

– 232 –
  • tion with the greyish green of some P. attenuata, it resulted in a dull, bronze-like brown. It was strongest in the lot from McCloud; the tops of these were a distinctly reddish purple hue. In the hybrid swarm it showed most vividly against some of the pale green tufts of primary leaves, and ranged from the merest tinge of pale mauve to strong purple.

    Some individuals must have been less exposed to cold than others, and this might account for some of the individual variation within each lot of P. attenuata or hybrids. Nevertheless, the difference between the species was striking, and the presence of purple in the hybrid lots can probably be ascribed to P. attenuata genes.

  • (15) Degree of Bud Formation

    Three subjective grades were used. In P. radiata the apices were mostly surmounted by tufts of pale green primary leaves, and only a few showed the earliest stages of transition from this to the brown, imbricated scale-leaves of a fully developed bud. The P. attenuata from Placerville all had compact, narrow, cylindrical buds, whitened by encrusting resin. The F1 hybrids all had terminal buds more or less developed, but these were fatter, less compact, and had scarcely any exuded resin. The McCloud lot had similar buds, but not on every tree. The Mexican lot showed very little bud formation, resembling P. radiata rather closely. The hybrid swarm had all degrees of bud formation, but the largest single category was the one coinciding with P. radiata.

Picture icon

Text-fig. 2.—Frequency distributions of hybrid indices, calculated from seven characters, for samples of six different populations. (Table II contains details.)

– 233 –

Analysis by Hybrid Indices

The characters chosen, and the method of allotting index values to them, are set out in Table II. The final results of the analysis are shown in Text-fig. 2.

From this it is apparent that the characters used did, in the aggregate, discriminate the two species well enough, particularly if the Mexican P. attenuata sample was excluded. They also placed the F1 sample in the intermediate position which was expected in theory.

Considering the hybrids from Ashley Forest, it may perhaps be assumed that their original P. attenuata parent was of Californian, not Mexican, lineage. If this be so, and if the known parent was an F1 hybrid, the frequency distribution suggests that most of them were the result of back-crossing to P. radiata, which was plentiful nearby. At the same time, the tail of the histogram extending towards the P. attenuata end of the scale suggests that one or more of the following also occurred:

  • (a) back-crossing to P. attenuata,

  • (b) selfing,

  • (c) pollination by one or more other hybrids.

Of these possibilities, the first can almost certainly be ruled out, because, as far as is known, P. attenuata never existed anywhere near the known parent. The others must be accepted, because there is no known intrinsic barrier to selfing, and because there is good evidence of other hybrids in the vicinity.

On the other hand, if the known parent was a hybrid of a later generation, e.g., an F2, and such that genes from P. attenuata outnumbered those from P. radiata in its genotype, its progeny might perhaps all have arisen from back-crossing to P. radiata. From a study of the phenotype and circumstances of the known parent, it was concluded that this latter interpretation was nearer the truth. Even so, some of the effective pollen grains may have been of hybrid origin.

Character–Combinations in the Hybrid Swarm

The new combinations of characters in the progeny from Ashley Forest were perhaps more diverse than might have been expected, but a general trend from the P. radiata complex towards the P. attenuata complex was discernible (Text-figs. 3 and 4).

By the principle of extrapolated correlates (Anderson, 1949: 92–101) one may visualize what the original P. attenuata parent would have been like at the same age, and under the same conditions of cultivation, as the others. It is pictured as similar to those shown in Plate 19, Figs. 8 and 9, with its foliage yellowish green and strongly tinged with purple, its young shoots yellow-brown and shining, its scale-leaves very dark-edged, and its stem apex encased in a perfect bud. The individual thus conceived would have clear affinities to the samples from Placerville and McCloud, but its foliage would be like that of the Ensenada sample. It would differ from all three in its very long and patent branches.


This study has its weaknesses, the worst of which probably lie in the choice of certain characters for the polygraphic analysis and for the analysis by hybrid indices. Nevertheless, it is presented for the following reasons:

  • (1) It shows that the application of these analytical techniques has been extended to unusual material, inasmuch as two-year-old trees are extremely young and show only some of the vegetative characters, and none of the reproductive characters, of maturity.

– 234 –
  • Picture icon

    Text-fig. 3.—Scatter-diagram of 35 individuals of P. radiata and 20 of P. attenuata (10 from Placerville and 10 from McCloud). The P. radiata sample forms the lower left-hand cluster; the P. attenuata sample forms the upper right-hand cluster.

– 235 –
  • (2) It supports the recent claim by Anderson (1954: 100–01) that these methods can give good results with small samples.

  • (3) It shows that differences between P. attenuata and P. radiata are manifested at a very early age.

  • (4) It affords satisfactory evidence that the progeny of the Ashley tree consisted of a swarm of hybrids between P. attenuata and P. radiata.

  • (5) It suggests that P. attenuata, in its natural state, probably has a morphological and genetic variation which is correlated with its wide territorial distribution. A knowledge of this variation might one day have an important application in the culture of P. radiata.


It is a pleasure to record the kindness of Dr. J. W. Duffield, who sent the seeds from California, and of Dr. F. I. Righter, who supplied more information about their origin.

Picture icon

Text-fig. 4.—Scatter-diagram of 50 individuals, which resulted from open pollination of a single hybrid. The symbols were constructed in the same way as those in Text-fig. 3. The dotted lines represent the limits of the samples shown in Text-fig. 3.

– 236 –


Anderson, E., 1949. Introgressive Hybridization, 109 pp. New York: John Wiley and Sons.

——1954. Efficient and inefficient methods of measuring specific differences. In Statistics and Mathematics in Biology, 93–106. Edited by Kempthorne, Bancroft, Gowen and Lush. Ames, Iowa: The Iowa State College Press.

M. H. Bannister, M.Sc.

Forest Research Institute,