Studies in Pimelea I—The Breeding System

Transactions and Proceedings of the Royal Society of New Zealand, Volume 88, 1960-61, Page 29

Studies in Pimelea I—The Breeding System By C. J. Burrows, Botany Department, University of Canterbury. [Read by title before Canterbury Branch, December 3, 1988; received by the Editor, March 11, 1959.] Abstract The paper describes the gynodioecious breeding system of natural populations of four species of the genus Pimelea1. 2 growing in a restricted area at Cass Some experimental work and observation of breeding behaviour in a year of one flowering season is discussed. The insects effecting pollination are noted. The breeding system is shown to make possible some factors of evolution, particularly hybridization, in the genus Pimelea. 1 A. detailed account of variation, hybridization and taxonomy in Pimelea is given in Burrows (1958). 2 The species of Pimelea at Cass are P. prostrata, Willd. Sp. Plant (1798), P. traversii, Hook. f., Handbook N.Z. Fl (1864), P. “shorb tussock” and P. “snow tussock”. The writer is not yet in a position to publish descriptions of the latter two species. Reference specimens are. P. “Short tussock”, No. 320 and P. “snow tussock”, No. 321 in the Botany Department, University of Canterbury herbarium, Cass Collection Hybrids are found, especially between P. “short tussock” and P. “snow tussock”, but also notably between P. “snow tussock”, and both P. travesii and P. prostrata. Iintrodution A Knowledge of the breeding system of a plant species is necessary in order to understand the patterns of group variation. On the breeding behaviour depends the genetic structure and ultimately the kind and degree of genetic, heritable variation. Darwin (1884) defines gynodioecy as the state where there is a group of plants having female and hermaphrodite individuals. The term gynodioecious is restricted to species of plants which maintain in their populations a high proportion of female plants, contributing significantly to the type of genetic structure of the population. If the selfed hermaphrodites set seed the recombination benefits of outcrossing are conferred while the certainty of inbreeding is retained. In many populations of hermaphrodit species, there is a small propotion of male sterility, resulting from abortion of pollen, which is not significant in consideration of the breeding system of a population (see Frankel, 1940). This is not to be regarded as gynodioecy. Lewis (1941) showed that where male sterility is due to a dominant or recessive gene the females cannot exist in wild populations unless more than twice as fertile as hermaphrodites (on the female side). This arises from the fact that hermaphrodites contribute potentially three times the amount of genic material that females do. Lewis, in this paper, explains the reason why, at that time (1941), cytoplasmic inheritance was the only explanation for the existence of gynodioecy in the natural populations of plants then investigated. In the case of cytoplasmic male sterility, only a slight advantage of the females is necessary to maintain them in the field, as opposed to the large advantage required when the process is controlled by nuclear genes. Although Correns (1928, ex Lewis 1941), Lewis (1941) and many other writers (mentioned by the latter) gave the gynodioecious breeding system a cytoplasmic.

basis, Lewis and Crowe (1956) have re-examined the field in terms of orthodox genetics since then. Some of the data published by Correns is found, in the light of work by East (1934, ex Lewis 1956), to be explicable on an orthodox genetic basis. C. & H. Yampolsky (1922) list only 97 species of 5 genera as being gynodioecious. They mention Pimelea as having 70–80 species, hermaphrodite polygamous or dioecious. Baker (1948) gave more than 70 species in 16 families in which functional female flowers were smaller than functional hermaphrodites. Many of these were gynodioecious. The Thymelaeaceae (to which Pimelea belongs) was not listed by him. Godley (1955 and 1957) described gynodioecy in Fuchsia, Cyathodes and Leucopogon in New Zealand. Gynodioecy in Pimelea In the most careful description of polymorphism in Pimelea flowers yet published, Thomson (1880) concluded that P. prostrata, “Though hermaphrodite in structure… is dioecious in function and will probaly tend rapidly to become so in structure as well”. Thomson described hermaphrodite forms where the style lengthened after the flowers opened until it extended beyond the anthers. Other dioecious forms had large male flowers on some individuals and smaller females on others. In the former, the short style, with tiny stigma, lengthened and protruded between the lobes of the flower tube, although the ovary then shrivelled. The female plants had flowers with a long style, a large capitate stigma and aborted anthers. P. longifolia was described by the same writer (Thomson, 1927) as being polygamous and dioecious. Cheeseman (1925) gives as a generic character for Pimelea that the flowers are hermaphrodite but functionally dioecious or occasionally polygamo-dioecious and that “The male flowers are the most numerous….” Cheeseman designates as polygamo-dioecious3. Polygamo-dioecious species are defined as having plants of different sexes, female and male, where one or other, or both, have few flowers of the opposite sex or hermaphrodite flowers, or both., erroneously, some 10 of the species described by him in the “Manual” including P. traversii and P. prostrata. The other species he describes are not classified, as to breeding system. Cheeseman (1914), figured four species and gave similar details. Parlane (1925) wrote about floral features in Pimelea but followed Cheeseman. Dr. E. Godley (pers. comm.) has found that P. longifolia and P. virgata are gynodioecious. He has noted that in the specimens used by Cheeseman for the illustrations of P. longifolia (sheets 5336.1 and 5336.2 in the Cheesemand herbarium) all hermaphrodites have protruding styles so that the male flowers described by Cheeseman are not present. It seems probale that Cheeseman had observed hermaphrodite flowers in different stages of development. As well as for the four species at Cass, the present writer has observed that P. aridula, P. sericeo-villosa, P. pseudo-lyalii and P. arenaria have female and hermaphrodite flowers on separate plants. It will be shown that on functional grounds also the Cass Pimeleas are gynodioecious. Australian and Tasmanian writers, Black (1952) and Cruickshank (1953) have not used the term gynodioecy for the type of breeding system in Australian Pimelea spp. The former says of the genus, that species are “bisexual or unisexual by abortion of stamens or ovary, often diocious”. The latter describes some species as dioecious. It appears that there is greater variation in types of breeding system in Australia than in New Zealand. Descriptions of some of the large number of Australian species of Pimelea (Bentham, 1873) leave little doubt, however, that gyno-

Fig. 1.—A, B, ♀ and ☿ flowers of P. “snow tussock” at the time of opening. Note position of ♀ ☿ stigmas with respect to ☿ anthers C, D, ♀ and ☿ flowers after 5 days, unfertilized. Note growth of styles.

dioecy is present Species with large hermaphrodite flowers and small females on separate plants exist, but other species described appeared to be wholly hermaphrodit. Flower Structure The flower of Thymelaeaceous plants is considered by Heinig (1951) to be an expanded torus, with the calyx at the end of the floral tube. The flower is best referred to as the flower tube, and this term will be used in future discussion of Pimelea flowers. A description of the flowers of one of the Pimelea species at Cass, in general terms, will cover all of them, except in gross dimensions, since little other difference is to be seen in flowers from different species. Flowers are terminal and grow erect in dense heads of 10–20 flowers. They are almost sessile on very short stalks, and at the base of the flowers is a dense growth of silky hair which remains after the fruit has fallen. The flower tube is cylindrical, with four calyx lobes at the distal end of a hollow tubular torus. The lobes are small and spreading when the flowers are open. Petals are absent, but the torus is petaloid. The upper side of the calyx lobes and inside of the floral tube is waxy white and usually glabrous (some hairs may be found, especially in P. traversii, inside the tube) and bears very small tubercles which give a shining appearance. The outer wall of the flower tube and underside of the calyx lobes are densely clothed with simple hairs. The base of the tube, surrounding the ovary, is red in P. prostrata, P. “snow tussock” and P. “short tussock” Especially in female flowers this area is often folded, and bulging. The hermaphrodite flower tube base bulges less and is narrower and longer than in females. The hermaphrodite flower tube base bulges less and is narrower and longer than in females. The flowers of P. prostrata, P. “short tussock” and P. “snow tussock” have a distinctly sweet scent, but the scent of P. traversii is negligible Nectar is produced in both females and hermaphrodites, apparently by morphologically unspecialised cells on the inner walls of the flower tube. It lies here, about half way up the tube or at the bottom of the tube. Measurements of the long axis of flower tubes in from 20–30 plants (average of 5 flowers per plant) in both sexes of each species at Cass are given below. In all cases the hermaphrodite flower tubes are larger than the female ones, although in any one species, a few individuals of either sex approach one another in size. Other structural differences between hermaphrodite and female are also Flower Measurements Species Mean Flower Tube Length Range Standard Deviation P. “snow tussock” ☿ 5.4mm 4.5–6.5mm 0.65 P. “snow tussock” ♀ 3.3mm 2.2–4.1mm 0.45 P. “short tussock” ☿ 4.4mm 3.5–5.2mm 0.49 P. “short tussock” ♀ 3.0mm 2.4–3.8mm 0.36 P. prostrata ☿ 4.0mm 2.6–4.9mm 0.43 P. prostrata ♀ 2.6mm 2.1–3.0mm 0.28 P. traversii ☿ 7.4mm 6.2–8.5mm 0.93 P. traversii ♀ 4.8mm 4.0–6.0mm 0.53

obvious. Stamens are two in number in hermaphrodit flowers and inserted just below the throat of the floral tube, opposit the two outer lobes. Filaments are slender and the anthers extend straight above the flower tube for 2–3 mm and face inward. Dehiscence of anthers in the hermaphrodite allows pollen to fall into the flower tube. In the female stamens are withered and filaments are shorter. Shrivelled pollen is found in a few cases. The ovary in both female and hermaphrodite is one-celled and the slender style rises from one side of it, near its summit. During the period after opening of the flower, the style, which lies to one side of the floral tube, begins to elongate. Eventu|ally, if the flower is unpollinated, it may extend from 1/2 - 3/4 of the length of the floral tube beyond the mouth of the tube. The stigma of the female flower is more than twice as wide as that of the hermaphrodite, and its simple papillae are also longer. They are twice the lengeth of those of the hermaphrodite and extend radially from the centre. (See Fig. 1.) In a few cases hermaphrodite plants carried a few female flowers in the heads of hermaphrodite flowers. Some such mixed heads had a few flowers of intermediate dimensions and with characters of both sexes. The upper part of the ovary is hairy. After fertilization the ovary begins to swell and in P. “snow tussock”, P. “short tussock” and P. prostrata the growth is sufficient to burst the persistent base of the flower tube. This serves as a loose outer coat, becoming glabrous brown and straite, but soon falls away. Although the tube does not appear to be articulate between base and upper tube, it often breaks here. Proportions of Plants of Different Sexes Present in Natural Populations. Species Number of ☿ plants Number of ♀ plants Ratio ☿ ♀ Cass P. “snow tussock” early in season 47 49 1 : 1 P. “-snow tussock” late in season 41 38 1 : 1 P. “short tussock” early in season 41 39 1 : 1 P. “short tussock” late in season 35 35 1 : 1 P. prostrata early in season 54 35 1.54 : 1 P. prostrata late in season 42 22 2 : 1 P. traversii early in season 32 25 1.28 : 1 P. traversii late in season 43 23 2 : 1 Erewhon P. “snow tussock” early in season 44 22 2 : 1 P. prostrata early in season 60 10 10 : 1 Macaulay River P. “snow tussock” 47 24 2 : 1 P. prostrata 55 18 3 : 1 Lake Tekapo P. “snow tussock” 37 33 1 : 1

In P. “snow tussock”, P. “short tussock” and P. prostrata only the base is left, but in P. traversii the whole hairy flower tube is left surrounding the ovary, although not adherent to it. The fruits of P. “snow tussock” and P. “short tussock” are oval, yellow, orange, or red and fleshy. They are somewhat flattened laterally, and the seed, black, hard and oval, is imbedded in the fleshy portion. Fruits of P. prostrata are similar in size, but snow white and fleshy. In these plants fruits are produced in large numbers in crowded clusters. Very few of the flowers of females do not produce fruit. In P. traversii there is a small green, hard fruit. Fruits are quickly shed when ripe from all four species. Natural Populations The table above shows the numbers of different sexes of plants counted in line transects at Cass in the season, summer 1957–58. The counts were made of every flowering plant on lines through each population. The ratios of the sexes one to another are also given. Included in the table also are figures for counts along the same transects at different parts of the flowering season and for some counts made of plants at Erewhon, in the Rangitata watershed in 1957, at the Macaulay River, and at Lake Tekapo in 1959. The plants counted at different parts of the flowering season will, in most cases, be different individuals, since any one plant flowers for a comparatively short period. The above figures should be compared with figures for absolute numbers of fruit set and for percentage fruit set in the same species at Cass. The latter were arrived at by counting between 20–30 flowers in heads of individual plants, then marking these and finally counting the number of fruit set for each of these. For 20 plants (10 of each sex) about 25 flowers each were counted in an area of a few square. Numbers of Fruit Set by Counted Flowers;. P. “snow tussock” ☿ ☿ Plant No. of Flowers Counted No. of Fruit Set Plant No. of Flowers Counted No. of Fruit Set 1 22 1 1 23 14 2 24 3 2 28 9 3 22 0 3 25 11 4 29 0 4 25 2 5 21 0 5 27 12 6 24 1 6 23 18 7 23 0 7 33 26 8 23 0 8 23 2 9 22 0 9 24 3 10 22 0 10 24 5 Total 232 5 Total 255 102

P. “short tussock” ☿ ☿ Plant No. of Flowers Counted No. of Fruit Set Plant No. of Flowers Counted No. of Fruit Set 1 25 0 1 24 15 2 24 1 2 25 11 3 24 0 3 25 23 4 24 1 4 24 23 5 22 1 5 21 19 6 24 0 6 27 20 7 24 3 7 22 16 8 24 2 8 23 20 9 27 1 9 25 24 10 24 0 10 23 17 Total 243 9 Total 239 188 P. prostrata ☿ ☿ Plant No. of Flowers Counted No. of Fruit Set Plant No. of Flowers Counted No. of Fruit Set 1 23 0 1 24 5 2 27 0 2 22 0 3 24 0 3 23 19 4 22 0 4 21 11 5 22 1 5 23 10 6 24 0 6 23 4 7 21 0 7 22 16 8 24 2 8 23 7 9 24 3 9 23 5 10 23 4 10 25 19 Total 234 10 Total 229 96

P. traversii ☿ ☿ Plant No. of Flowers Counted No. of Fruit Set Plant No. of Flowers Counted No. of Fruit Set 1 23 0 1 27 23 2 24 7 2 22 6 3 25 13 3 22 16 4 28 5 4 32 28 5 24 20 5 23 22 6 23 11 6 35 35 7 22 15 7 23 17 8 26 8 8 32 30 9 27 7 9 33 31 10 25 2 10 Total 247 88 Total 249 208 chains. The range of variation in fruit set from plant to plant both in females and hermaphrodites is of interest. Figures for P. prostrata, P. “snow tussock” and P. “short tussock” are fairly consistent, but those for P. traversii are more variable, as in seen in the table below. It is possible that error could arise through loss of some fruit before counting P. “snow tussock” in a shady gully may have been affected in this way since a few female plants had far fewer fruits than the expected number. This may account for some of the variation in female fruit set. For this latter group of plants two sets of calculations, one involving all the plants and the other excluding those with very low fruit set, are given in the percentage fruit set table. Percentage Fruit Set in Plants of Different Sexes Species Mean % Fruit Set Ratio ☿ ♀ ♀ ☿ P. “snow tussock” low figures included 40.0 2.15 16.7 : 1 P. “snow tussock” low figures excluded 56.25 2.15 26.1 : 1 P. “short tussock” 78.66 3.72 21.4 : 1 P. prostrata 41.92 4.27 9.8 : 1 P. traversii 83.54 40.64 2.0 : 1

The importance of these tables in showing that, firstly, the different Pimelea spp. vary considerably with respect to proportions of different sexes in the populations and secondly, population differences in the same species from area to area indicate that breeding system structure varies from population to population. The figures given by Godley (1955) for variation in proportion of females to hermaphrodites in Fuchsia from area to area show similar conditions. The differences from species to species are probably inherent in the genetics of the plants, but the population differences may be due to selective influences. An examination of these figures shows that at Cass both P. “snow tussock” and P. “short tussock” with 50% each of females and hermaphrodites in the population also have about 20 female to 1 hermaphrodite fruit set. This is important as a compensating factor, allowing the high proportion of females to be maintained in the field. On the other hand P. prostrata and P. traversii, with only about 30% of female plants in the population, have about 10 female to 1 hermaphrodite and 2 female to 1 hermaphrodite fruit set respectively. The higher proportion of hermaphrodite fruit set determines that fewer females may be retained in the population. It is to be expected that the Erewhon population of P. prostrata has yet a lower female fruit set. Low fruit set in hermaphrodites may perhaps be connected with incompatibility. Although these figures are indicative, and most useful in gauging the nature and extent of gynodioecy in Pimelea spp., the picture would be more complete with a study of the proportions of female to hermaphrodite seed which viable. This was not carried out owing to lack of time and difficulty of germinating seed. The differences in fruit set compared with percentage numbers of the different sexes in the field in P. prostrata and P. traversii indicate differences in the sex determination mechanisms. It will be noticed that the percentage of female flowered plants counted in the populations of P. prostrata and P. traversii early in their flowering seasons was greater than at a later stage. There may be some connection in this with hormone balance. The higher proportion of female flowers open early in the season could, however, if insect vectors are active, contribute to the high proportion of females in the population. The competition between insects for flowers might have some bearing on this. Crossing Experiments Some closely controlled experiments were carried out at Christchurch and at Cass, where plants of the four different species were used to assess relative amounts of fruit set by female and hermaphrodite plants upon artifical crossing. Selfing of hermophrodites and cross pollination of hermaphrodites and females were used in these intraspecific crosses. No adequate figures for statistical treatment were obtained and no seed could be germinated to test viability. All flower heads used were bagged prior to opening of flowers, and care was taken to avoid accidental crosses. In all cases females set fruit in a high percentage of flowers, but very few attempts at crossing and selfing of hermaphrodites were successful. Enough evidence was gathered to show that P. “snow tussock”, P. “short tussock” and P. prostrata (the two latter more rarely) can all set fruit through self pollination of hermaphrodite flowers, but it is obvious that this is less efficacious than cross pollination of females. Cross pollination of emasculated hermaphrodites also results in a low percentage of fruit set in P. “snow tussock”.. Upon selfing, only a few fruit per head were set at Christchurch on P. “snow tussock” hermaphrodites but P. traversii hermaphrodites at Cass showed higher self fertility (8–10%). Artificial crosses with hermaphrodites thus resulted in fruit set comparable with that in natural conditions. Dr. E. Godley (pers comm.) has verified self fertility of hermaphrodites for P. traversii.

These experiments in Pimelea show that, in fact, not much more than half the amount of genic material supplied to a new generation comes from the hermaphrodites in P. “short tussock” and P. “snow tussock”, since hermaphrodite fruit set, whether due to selfing, or cross fertilization, is very low in them. Hermaphrodite fruit set is only about 4–5% of the total fruit set by females in natural populations of these two taxa. In P. prostrata the percentage fruit set of hermaphrodites as compared with females is also low—about 10%, but in P. traversii it is much higher —just less than 50%. The systems of inheritance of sex may differ between P. traversii and the other species but it is possible for them all to have a genetic basis. Cruickshank (1953) discussing Tasmanian species of Pimelea stated that “All species examined have fertile pollen, and there is no evidence of any apomixis in the genus.” The specimens from Cass were tested for apomixis by bagging (and emasculation of hermaphrodites) but no fruit was set by unpollinated flowers. All the species had apparently fertile pollen. Examination was made of pollen stained with cotton blue in lactophenol, and in each species the grains were full. Only a small percentage of pollen was deformed, abnormally small or empty. The Pollen Vectors Numbers of insects were captured on flowers of the various species of Pimeleas at Cass Some were sent to the Cawthron Institute for identification, and Dr. E. S. Gourlay determined some of these. Other insects were determined by Dr. R. Pilgrim of the Zoology Department, University of Canterbury. The writer made further determinations by comparison of insects with already identified specimens. The details about insects are tabulated below. Most of the flies, the little purple butterfly, Crysophanus boldenarum and the small bee Halictus huttonii are ubiquitous in the short tussock areas at Cass. On fine days when Pimeleas are flowering, some flies and the bee in particular are present in large numbers, flying from flower to flower. Their main purpose is to suck nectar, and pollination is achieved incidentally to this, except in the case of occasional pollen collection by Halictus. Close observation showed that flies and bees had pollen adhering to body hair and on the head. Occasional individuals of Halictus collected pollen on the hind tibia. Although pollen was seen on bees and flies none of this was examined closely. The foraging range of individual insect species was also not examined closely but some general observations were made. Of the more frequent visitors, most of the Dipterans seemed fairly restricted in range. Crysophanus boldenarum, however, although it remained in one area for a time, was able to fly long distances. Halictus huttonii has its home in clay patches and banks, and is a solitary bee, but is quick flying and appeared to be the most widespread pollinating agent. Halictus and some of the flies by virtue of their presence in numbers are the most important vectors. Hundreds of bees and flies are found within areas of a few hundred square yards. Up to a dozen bees could be seen on one plant or small group of plants and half a dozen flies would be present in the same area. The insects moved quickly from flower to flower. Nectar, which accumulates at or near the bottom of the floral tubes was eagerly sought. Scent is most noticeable on sunny days about the plants (except P. traversii) and combined with the white shining calyx lobes serves as a daytime attractant to butterflies, bees and flies. Apart from sweet scent and colour the showiness of flower heads is important. Although individual flowers are small, their presence in crowded heads with up to 1,000 flowers per plant ensures that a showy mass of flowers is presented to insect visitors. The flower heads of P. traversii are extremely showy. In visiting Pimelea flowers there was no distinction by the insects between female or hermaphrodite plants growing side by side. Several plants of P. “snow tussock”, P. “short tussock” and hybrids (between them) were seen to be visited.

Insect Visitors to Pimelea Spp. at Cass. Order. Family Genus and Species. Determination. Pimeleas Visited. Description. 1. Hymenoptera Andrenidae Halictus huttonii Cam. E.S.G. P, A, L, T, and hybrids Small black bee. 2. " ? sp. indet. T. Larger bee, collects pollen. 3. Diptera Tachinidae Erythronychia velutina Mall E.S.G. P, A. Large fly. 4. " " Zealandotachina varipes varipes Mall. E.S.G. P. A, L. and hybrids small nondescipt fly. 5. " " Heteria sp. E.S.G. P. samall haisrsy fly. 6. " " cf. Protohystrica alcis R.P. T. Large orange bodied fly. 7. " " sp. indet. A. Small grey fly. 8. " Muscidae Calliphora laemica R.P. A. Large brown blowfly. 9. " " sp. indet. A, T. Medium sized grey fly. Few wing marks. 10. " Syrphidae Syrphus novae-zealandicus R.P. T. Hover fly. 11 " Anthomyidae sp. indet. P. Fly with mottled abdomen and wings 12. " ? sp. indet. T. Large brownish fly. Resembled hover fly. Cream body marking. 13. " ? sp. indet. A, P. Very small blueish fly. 14. " ? sp. indet. A. Rusty brown hairy fly. 15. Lepidoptera ? Crysophanus boldenarum C.J.B. P, A, L. Small purple-winged butterfly. 16. " C. sallustius C.J.B. T. Larger orange-winged butterfly. 17. " sp. indet. P. Small brownish moth. 18. " sp. indet. A. Fawn winged moth. 19. Coleoptera Cerambycidae Zorion cf. guttigerum C.J.B. T. Blue-black beetle with 2 orange spots. 20. " ? sp. indet. T. Small blue-black beetle. 21. " Curculionidae sp. indet. T. Large light brown weevil. 22. Hemiptera Lygaeidae sp. indet. T. Small Hemipteran insect. 23. ? ? sp. indet. L, X, A Brown long bodied insect. hybrid E.S.G.—E. S. Gourlay. R.P.—R. Pilgrim. C.J.B.—the writer. P.—P. prostrata. A.—P. “short tussock”. L.—P. “snow tussock”. T.—P. traversn.

indiscriminately by the same bee or fly. At this time, the same vectors were observed visiting P. prostrata plants. Few insects were seen on any one P. traversii plant. While it was noticed that flies and Halictus worked systematically over a flower head in the othe species, in P. traversii, Crysophanus sallustius was seen to be present for long periods and is an important vector. Most P. traversii individuals flower later than the other Cass species, and most of the individuals grow amongst thick scrub, so that this may account for different animal populations on P. traversii. However, a number of single plants of P. traversii are mixed with all the other species in some areas and flowering times overlap. Similarly P. prostrata, P. “snow tussock” and P. “short tussock” are all juxtaposed in some places and these all have overlapping flowering times. In any one day the P. “snow tussock” plants, growing in shady places, are visited for shorter times by the sun-loving flies, butterflies and bees than are P. prostrata and P. “short tussock” in open sunlight. Some P. “snow tussock” plants grow on fans and exposed places, however. No night flying insects have been taken into account in describing pollinating agents, but many moths are found in the tussock country. The scented white flowers of P. prostrata, P. “short tussock” and P. “snow tussock” no doubt attract night flying insects. Since distances between different Pimelea species, flowering at the same time, are not great in some places at Cass (some are only feet apart), the physical possibility of cross pollination between species as well as outbreeding within species is important in considering hybridization and variation. The non-selectiveness of unspecialized vectors may help to account not only for similarity in flowers in different Pimelea species but for the amount of hybridization taking place. (See Thomson, 1880, 1927). A. number of the same insects have, in fact, been captured on as widely different genera as Aciphylla, Leptospermum, Leucopogon, Brachycome, Celmisia and Hypochaeris, but when Pimeleas were in full flower other profuse flowerers like Leptospermum were not yet out. It was not ascertained whether one individual Halictus restricted itself to one genus or not. The effect of heavy rainfall on pollination and fruit set is difficult to gauge. In the 1957-58 summer it appeared to have had considerable significance in decrease of total fruit set. The flowers of one head may be opening over a period of a little more than a week, and the flowers of each plant over a two-week period. They soon wither and die when fruit begins to form or if they are unpollinated. Most plants of any one taxon in a given area flower together. If some catastrophe occurs this may result in a reduction in numbers of fruit set. As a general observation, it was noticed that much more fruit were set, especially in P. “short tussock” and hybrids in the summer of 1956–57. P. traversii and P. prostrata were less affected. Heavy continuous rain affects not only the ability of insects to fly and seek nectar, but it may drown out such ground dwellers as Halictus and also fills floral tubes with water, washing pollen off anthers. The most favourable weather for animal activity and for optimum attractiveness to animals of Pimelea flowers is the hot, dry, sunny weather usually experienced at flowering time. Relations of Floral Parts at Pollination Growth of floral parts is important when detail of the breeding system is to be considered. In both females and hermaphrodites it was noticed that after flower opening, the styles elongated and were conspicuously exserted. At first, this was thought to indicate that hermophrodites were protandrous, since anthers dehisced as soon as flowers were open. However, Dr. E. Godley (pers. comm.) noticed that only in unpollinated flowers were styles strongly exserted. A series of tests, consisting of bagging, emasculating (hermaphrodites) and pollinating flowers, showed that this was so. From the time of flower opening both female and hermaphrodite.

flowers were examined as to (a) original position of the stigma with respect to the mouth of the floral tube; (b) final position of the stigma; (c) position of stigma at which pollination is successful in initiating fertilization; (d) the degree of inhibition of style growth due to pollination. Careful observations were necessary for this part of the study. Pollination was found to inhibit most of the growth of styles in both females and hermaphrodite flowers, and furthermore the stigmas were found to be receptive to pollen (resulting in fruit set) over a limited period. This extends for 1–3 days after the flowers open and they begin to wither within a week. When female flowers open the stigma is just above the mouth of the flower tube, but in the hermaphrodite it is still below the mouth of the tube (about a third of the way down it). Growth of the style is continuous after the opening of flowers unless pollination occurs, when the growth is slowed down. In the hermaphrodite during the 1–3 days when successful pollination can occur the stigma is below the level of the mouth of the flower tube. Once the stigma is exserted beyond this, fertilization cannot occur although pollination still inhibits most of the growth. In some cases it was found that pollen from one species is capable of halting style growth of another species. P. prostrata pollen on P. “snow tussock” stigmas and P. “snow tussock” pollen on P. prostrata stigmas was found to act in this way. This inhibition of style growth holds for both female and hermaphrodite flowers but fertilization of the latter is more rare. If unpollinated, the style of hermaphrodite flowers may eventually grow to nearly twice the length of the floral tube. Female styles rarely grow to more than one and a-half times its length in the same circumstances. Because of this it is possible to see in the field which flowers are pollinated although this is not a direct measure of fertilization and fruit set. A high proportion of hermaphrodite flowers in the field were noticed to be unpollinated. Growth of the style is important mainly in connection with selective influences in the breeding system. Because of differences in stigma and stamens, both in gross structure and, in their position within the floral tube when mature and because of the short time of their receptivity to pollen, it may be shown that a situation favouring cross pollination is present in Pimelea. The floral tubes in female and hermaphrodite flowers are of differnt length and styles of about the same length when the flowers open. (Fig. 1.) The implications of this, as concerns the insects involved and cross pollination are as follows. The insects visiting female and hermaphrodite flowers when both are functional will be more effective in corss pollination between sexes than in self pollination of hermaphrodites (although some self pollination of or crossing between hermaphrodites may be brought about by insects). This situation is dependent on the hermaphrodite stigmas being receptive to pollen only when low down in the floral tube. Insects will not normally carry pollen to these stigmas. They are not receptive when they are exserted, which occurs when they grow older (if not pollinated). Position of receptive stigmas in females relative to anther position in hermaphrodites, together with larger papillae on female stigmas, which pollen adheres to more readily, encourages outcrossing between sexes. Crude tests consisting of rubbing a clump of pollen grains on a needle against both female and hermaphrodite stigmas showed that pollen stuck to the larger female papillae more easily. Natural selfing of hermaphrodites without aid of insects, can and does occur as was shown experimentally, but this is rare. From the flower count and fruit set data it is seen, however, that a significant small percentage of hermaphrodite flowers set fruit in all the taxa. This must result largely from cross fertilization, and ensures that a number of fruit are set in hermaphrodites. Presumably there is some physiological barrier, genetically controlled, which prevents a very high percentage of hermaphrodite fruit set. This needs further investigation, although genetic experiment in Pimelea is a long term project. The females, however, set fruit freely in all species.

It is a possibility that some pollination is brought about by wind in Pimelea, but pollen is heavy and shedding of pollen from inward facing anthers, unless it is carried away by insects, usually results in the bottom of the hermaphrodite flower tube being filled with pollen grains. Since the receptive stigma is below the level of the anthers, although usually to one side of the floral tube, there is a strong chance of pollen falling on it. This does not often occur or the resulting pollination is not effective, as is shown by the large number of exserted styles seen in the field in hermaphrodites. It has been shown previously that effective pollination inhibits style growth in both female and hermaphrodite flowers. In the field, fertilized flowers with non exserted styles were seen to bear pollen on their stigmas and occasionally pollen could be seen on exserted styles. Modifications of Sex Expression Some plants in cultivation, and others in the wild populations at Cass and Macaulay River were noticed to bear female flowers in the same heads as hermaphrodites. These do not conform strictly to the implications of the term gynodioecy but are gynomonoecious. At Christchurch P. “snow tussock” and P. prostrata plants, members of clones, bore mixed flowers, although in each case the rest of the clone was made up of purely hermaphrodite plants. At Cass a few plants of all the species had the same condition, but P. traversii always had more female flowers on hermaphrodite plants than the others. The numbers of female flowers could have had some effect on the genetic structure of the species in P. traversii (since fruit set was established for female flowers in mixed heads) but probably is not important for the other species. In several cases, P. traversii plants marked as females early in the season, with hundreds of female flowers, had, later in the season, hundreds of hermaphrodite flowers. Many other plants had numbers of female flowers in hermaphrodite heads. Production of a few female flowers en bloc at the beginning of flowering by an otherwise hermaphrodite plant was the most usual situation. The early production of female flowers on these plants—they are usually the outermost flowers and first to open, suggests difference in environmental treatment. The differences must, however, be slight in nature. Heslop-Harrison (1957) puts the cause of such phenomena in flowering plants to variations in hormone supply. Checks of the numbers of different sexes of P. “short tussock” and P. “snow tussock” at Cass in both 1956–57 and 1957–58 summers revealed that there was no change in relative numbers of sexes from year to year. It seems that, although immediate control of sex may be hormonal, the ultimate causes of sex difference are genetic. In addition to production of clearly female and hermaphrodite flowers in the same head, some of the Christchurch and Cass plants were noticed to have “intersexed” flowers. Most of the female flowers on hermaphrodite plants had typical female dimensions. A few such flowers had one anther and small stigma papillae, but female dimensions. Others were larger than females, although smaller than hermaphrodites and some had large and some small papillae. Further female plants with a few flowers having one or two anthers were seen. A very few plants in a population were hermaphrodites with very small flowers or very large flowered hermaphrodites. Again this tends to bear out the idea that sex is controlled, immediately, by a hormonal balance mechanism. Sex Determination The determination of sex in Pimelea is of some interest. It is highly probable that size of flower and other polymorphic variations are linked to the cause of sex difference. Some meagre data about time of differentiation of female flowers show that pollen is produced but degenerates in the anthers of at least some flowers.

Chromosome counts of stages of meiosis could be made from anthers of some female flowers of P. traversii and, especially in a few flowers of P. “short tussock” and P. “snow tussock”, pollen was noticed during the very early development of female buds in marked plants. By the time buds were ready to open their anthers were shrivelled and pollen was unrecognizable. In one female P. “short tussock” plant with mature flowers, the anthers, indehiscent and shrivelled, contained collapsed pollen. Frankel (1940) found this phenomenon in Hebe townsoni. Some of the meiotic figures in plants at Cass (the sexes of the parent plants were not determined) appeared to show one set of unpaired chromosomes during second metaphase. These may have been females with degeneration setting in. In Pimelea some emasculated hermaphrodite flowers showed slight reduction in size from normal. It appears possible that corolla growth is influenced by a hormone produced in the anthers and released post-meiotically. Determination of corolla lengths probably occurs at a stage later than that of the determination of sex of the flower. A paper by Plack (1957) deals with a very similar situation in a sexually dimorphic Labiate, Glechoma hederacea. In this normally gynodioecious species she found that some individuals produce female flowers only, others are gynomonoecious, and the proportion of female to hermaphrodite flowers varies with the stock and the season. This parallels the situation in P. traversii at Cass in all respects. Glechoma is four anthered. Flowers intermediate in size between the small females and larger hermaphrodites had one, two or three anthers. Artificial emasculation of hermaphrodites in bud showed that corolla size was reduced. Plack thought that corolla growth is influenced by a hormone produced in the anthers and released after meiosis since the critical time for growth of corolla (or lack of it) was at a certain corolla length. There appears to be a close similarity between Pimelea and Glechoma in these respects. In a more recent paper (Plack, 1958) it is shown that treatment with Gibberellin caused growth of small female corollas in Glechoma, to the larger hermaphrodite size. Baker (1948) states that the wide range of dicotyledons in which small female flowers and larger hermaphrodites are found indicates a common factor in their origin, and there is a strong suggestion from the above evidence that hormone balance controls flower size. A considerable body of information is summarized by Heslop Harrison (1957) who wrote of experimental modification of sex expression. It has been shown in certain plants that various factors of environment may affect development of flowers (with consistent experimental results for either monoecious, dioecious or hermaphrodite plants) causing variations in sex expression. These variations are considered to be due to hormonal causes, the environmental factors affecting production or movement of hormones which control sex. Heslop Harrison suggests that, as flowers are initiated at the apices, the auxin levels favouring optimum development of one sex suppress the other. In a normal environment the mechanism of sex inheritance establishes in some individuals one developmental path and in some the other. Baker (1948) states that “Because functional pistillate flowers may be distributed in a gynodioecious or gynomonoecious manner in the same species, abortion of anthers cannot always be determined by direct genetic means…”. These ideas do not, however, necessarily conflict with the concept of genetically controlled gynodioecy. Most individuals in populations of Pimelea are clearly of one sex or the other and remain so from year to year.

Discussion The Pimelea species at Cass, and others in New Zealand are gynodioecious both in structure and function, although some variations in sex expression are extant, as they are in many plant and animal populations. Some genetic experiment is needed to establish the basis for control of sex determination in Pimelea and to discover reasons for variation between different populations in numbers of the sexes. In wild populations the presence of relative proportions of one sex to another is interpreted by Heslop-Harrison (1957) as being due to the interaction of environment with the hormone system of the plants. In Pimelea maintenance of a high proportion of females benefits the species by ensuring outcrossing. This is enabled by activity of animal pollinating agents and the dimorphism of the sexes. Differences in pollinating species in different areas may contribute to variations in proportions of females to hermaphrodites. Ford (1957) in discussion of “intersexes” and the heterostyly of Primula shows how the control of these conditions is genetic and changes are induced by selective influences. Gynodioecy is in the main an outbreeding system. The higher the proportion of females to hermaphrodites, the greater the amount of outcrossing which is possible. Thus P. “short tussock” and P. “snow tussock” each with 50% of their populations females, at Cass, potentially are able to exchange genes most readily within or between species. The consequences of this, especially with regard to the plasticity of and hybridization between the Pimelea species are most interesting. The probable selective action of flower structure and insect activity in maintaining this situation are supplemented by the low fruit set in hermaphrodite plants as compared with females. However, a little inbreeding can take place, and this could also be of benefit. In P. prostrata and P. traversii the proportion of females is not as high but must also have an important bearing on their variation and ability to hybridize. It is clear, however, that the amount of hybridization between them and other species is more limited than between P. “snow tussock” and other species. They are also less variable throughout their range than P. “snow tussock”. The outbreeding of the latter thus contributes to its plasticity and to the amount of hybridization occurring between it and other species, including P. aridula and P. sericeo-villosa at Cass and elsewhere. Some of this hybridization (observed in terms of variations in morphology and supported by experimental work) is very far-reaching, and probably of the type described by Anderson (1953) and others, as introgression. Widespread hybridization in the genus Pimelea may be attributed in a large measure to gynodioecy in many of the species. Acknowledgments The writer is indebted to Professor W. R. Philipson for much assistance and encouragement and to Dr. E. Godley for his most valuable advice and criticism. The information which he made available was extremely helpful. Many other people assisted in various ways and their help is gratefully acknowledged. References Anderson, E., 1953. Introgressive Hybridization. Biol. Revs., 28, p. 280. Baker, H. G., 1948. Corolla Size in Gynodioecious and Gynomonoecious Species of Flowering Plants, Part II. Proc. Leeds Phil. Lit. Soc., 5, p. 136. Bentham, G. and von Mueller, F., 1873. Flora Australiensis, London. Black, J. M., 1952. The Flora of South Australia, Part III. p. 592. Burrows, C. J., 1958. Variation in Some Species of the Genus Pimelea. M.Sc. Thesis, University of Canterbury Library.

Cheeseman, T. F., 1914. Illustrations of the New Zealand Flora, 2. — 1925. Manual of the N.Z. Flora. Cruickshank, R. H., 1953. Chromosome Numbers in the Genus Pimelea. Proc. Roy. Soc. Tasmania, 87, p. 13. Darwin, C., 1884. Different Forms of Flowers on Plants of the Same Species. London. Ford, E. B., 1957. Polymorphism in Plants, Animals and Man. Nature, 80, No. 4598, p. 1317. Frankel, O. H., 1940. Studies in Hebe, II: The Significance of Male Sterility in the Genetic Ststem. J. of Genetics, 40, p. 171. Godley, E. J., 1955. Breeding Systems in New Zealand Plants, 1: Fuchsia. Ann. Bot., 19 n.s., p. 549. — 1957. Unisexual Flowers in the Ericales. Nature, 180, No. 4580, p. 284. Heinig, K., 1951. Studies in the Floral Morphology of the Thymelaeaceae. Am. J. Bot., 38, p. 113. Heslop-Harrison, J., 1957. The Experimental Modification of Sex Expression in Flowering Plants. Biol. Revs., 32, p. 38. Lewis, D., 1941. Male Sterility in Natural Populations of Hermaphrodite Plants. New Phytol., 40, p. 56. — and Crowe, K., 1956. The Genetics and Evolution of Gynodioecy. Evolution, 10, p. 15. Parlane, B. J., 1924. The Pimeleas of the Cass Valley, Canterbury. Unpublished M.Sc. Thesis. Plack, A., 1957. Sexual Dimorphism in the Labiates. Nature, 180, No. 4596. — 1958. Effect of Gibberellic Acid on Corolla Size. Nature, 182, No. 4635. Thomson, G., 1880. On the Fertilization of New Zealand Flowering Plants. Trans. N.Z. Inst., 13, p. 271. — 1927. The Pollination of New Zealand Flowers by Birds and Insects. Trans. N.Z. Inst., 57, p. 106. Yampolsky, C. and H., 1922. Distribution of Sex Forms in the Phanerogamic Flora. Bibliotheca Genetica 3, p. C. J. Burrows, M.Sc., Botany Department, University of Canterbury, Christchurch.