Publicly accessible
URL: http://nzetc.victoria.ac.nz/collections.html
copyright 2006, by Victoria University of Wellington
All unambiguous end-of-line hyphens have been removed and the trailing part of a word has been joined to the preceding line, except in the case of those words that break over a page. Every effort has been made to preserve the Māori macron using unicode.
Some keywords in the header are a local Electronic Text Center scheme to aid in establishing analytical groupings.
Tuatara aims to stimulate and widen interest in the natural sciences in New Zealand, by publishing articles which (a), review recent advances of broad interest; or (b), give clear, illustrated, and readily understood keys to the identification of New Zealand plants and animals; or (c), relate New Zealand biological problems to a broader Pacific or Southern Hemisphere context. Authors are asked to explain any special terminology required by their topic. Address for contributions: Editor of Tuatara, c/o. Victoria University of Wellington, Box 196, Wellington, New Zealand. Enquiries about subscriptions or advertising should be sent to: Business Manager of Tuatara, c/o. Victoria University of Wellington, Box 196, Wellington, New Zealand.
(This issue edited by J. A. F. Garrick)
is the journal of the Biological Society, Victoria University of Wellington, New Zealand, and is published three times a year. Joint Editors:
Insects are primarily terrestrial animals but one of the conspicuous features of this class is its adaptability to a wide range of habitats and even within New Zealand there are many aquatic and semiaquatic forms (Wise 1965). Among them are the Dytiscidae, the carnivorous water beetles, the first New Zealand species of which was recorded in 1846. By 1893 fourteen species were described and by 1920 this country was attributed with nineteen species. The early lists however relied on research carried out during the last century which was, of necessity, based on very few specimens and consequently it failed to take into account variation within the species. Today, with improved techniques and adequate series of specimens, specific variability can be assessed and synonymy in the classification of the family can be established. A full systematic revision has been carried out by the present author (Ordish 1966), and the reader is referred to it for detailed descriptions, anatomical details, and discussions of synonymy and distribution. It is the aim of this paper to provide a ready means of identification of the New Zealand Dytiscidae, to illustrate the morphology referred to in the key, and to present some notes on the general biology of this family.
The family Dytiscidae can be defined as aquatic, carnivorous coleoptera, principally of the temperate regions, which are oval, evenly rounded and streamlined in body form. Their eyes are prominent and undivided, their antennae simple, eleven-segmented and without pubescence. The thorax bears a ventral prosternal
The most distinctive feature of this family of beetles is the ged plate-like post coxa which occupies a large area of the ventral surface leaving the trochanter as the first movable segment of the leg. This, together with the undivided eyes and eleven-segmented antennae, serves to differentiate this from other families of the Coleoptera.
As will be seen from the foregoing key, identification at species level is often facilitated by determining the sex of the specimen. This is readily accomplished in the genera Homoeodytes, Dytiscus, Rhantus, Lancetes and Copelatus because in these the males have the three basal segments of the anterior tarsus much broadened so that together they form a disc, typically bearing suction cups (palettes) on the under surface. The anterior tarsi of the females of these genera are narrow and cylindrical. Males of Hyphydrus elegans may be recognised by their black anterior tarsal segments, those of the female being light brown. Within New Zealand, males of the genus Antiporus are readily recognised by the presence of an inwardly directed triangular protruberance on the hind femur, but separation of the sexes of members of Liodessus and Huxelhydrus is not so easy. Accurate determination can be done only by dissection but usually males of these two genera are shiny in appearance and narrower in the body, while the females are broader, and the upper surface is dulled by microsculpture. Sexing of the members of these two genera, however, is not necessary for identification.
The key also makes reference to the epipleural pit, a term which requires elaboration. The epipleura is the deflexed lateral
Hyphydrus the basal portion of the epipleura is separated by a conspicuous transverse ridge and the basal portion thus confined is termed the epipleural pit.
Morphologically the Dytiscidae show a close relationship with the carnivorous ground beetles, the Carabidae, and possibly they have evolved from this group. While the similarities are quite apparent, so too are the modifications to meet a rather demanding habitat. These are apparent both in structure and in behaviour. For most aquatic carnivorous animals there is survival value in effective swimming and in the dytiscids this is achieved by the use of greatly modified hind legs which serve as oars. Complementing this is a streamlining of body form to give a continuous oval outline; firm application and interlocking of the head, thorax and abdomen to give rigidity, and a smooth surface to provide minimum resistance to movement. Like the ground beetles they are nocturnal and remain hidden during the day. They are, however, air breathing and this requires periodic visits to the surface. Most dytiscids can remain submerged for about twenty minutes because they retain a considerable quantity of air beneath the wing cases in addition to the bubble of air often seen attached to the end of the abdomen. Replenishment at the surface is very rapid and usually takes about three seconds.
Dytiscid eggs are laid in the water and the larvae undergo four moults during growth, and feed on a variety of aquatic life including smaller examples of their own kind. Those larvae whose biology has been studied have been found to have hollow mandibles through which digestive enzymes are injected into prey thus providing an external digestion. Later the resultant fluid is pumped back through the mandibles into the alimentary canal. The available evidence would suggest that a given species has a favoured food species but field observations of this type are difficult. In captivity the larvae of Homoeodytes hookeri show a marked preference for Anisops assimilis, a backswimmer, when offered a range of food, and consume an average of one insect every twenty four hours. The mature larva leaves the water and burrows into the ground where it excavates an oval pupal chamber. The emerging adult returns to the water and may remain there, although for many species dispersal by nocturnal flight is not unusual.
In New Zealand there is much work to be done on the distribution of this group but the available evidence would suggest that many species are widespread probably because they have retained the power of flight. Some species, however, are restricted in their range and it is the author's intention to continue mapping their
Habitat preference, of course, varies from species to species, but generally speaking the still pond is the favoured site although some species are equally at home in the quieter backwaters of streams. The greatest numbers, however, occur in temporary and permanent ponds that have the organic content to support the animal life on which the dytiscids feed. Lakes and streams on the other hand usually have a modest quota of beetles. The significance of pH values has yet to be investigated but alkaline conditions appear to be the most favourable.
In contrast to the moderate flexibility of habitat of the New Zealand water beetles, Huxelhydrus syntheticus is essentially a river-bed species inhabiting the margin of larger rivers, and in particular the pools in marginal shingle. Copelatus australis is known only from Ruatoria on the east coast and Mokohinau Island in the Homoeodytes and Antiporus, but other possibilities cannot be assessed until more is known of this group in the Pacific region.
Based on a paper read at the Ninth N.Z. Science Congress, Wellington, 1960. Diagrams prepared at that time by
Hebe is credited with more species than any other genus of New Zealand plants, and is reputed to be almost the most difficult. The simple, mostly entire leaves offer few obvious characters to distinguish one species from another, and the small flowers are a good deal alike and are nearly all white. Yet the species differ far more than do sheep on the hill which the farmer recognises individually! It is a matter of training one's eye by looking, consciously or unconsciously, at those features that offer the strongest and most reliable contrasts.
The present notes add little to the scheme already available (Moore in Allan Fl. N.Z. 1961, pp. 885-887) but some learners may find pictorial representation helpful. Diagrams give a chance also to suggest possible relationships. The comparisons are extended beyond Hebe by including parallel information about Pygmea and Parahebe, two genera which, like Hebe, have been split off from the older genus, Veronica.
The seventy-nine New Zealand species of Hebe are arranged in ten groups, the largest group including 21 species, the smallest composed of a single one. Members of each group share certain common features, or vary within a limited range, and in assigning a specimen to its proper group one learns to recognise contrasts within the genus. A summary is presented here in the form of four sets of diagrams, each set with twelve insets, ten of them representing the groups of Hebe, the other two showing Pygmea and Parahebe.
The first set of diagrams (Fig. 1) gives the names of the groups and the number of species in each, and indicates the broad distributional pattern. Approximate altitudinal range is suggested by the terms low, mid, and high. A square inset shows that the group lacks North Island representatives, while an eight-sided one shows that at least some of the included species are found in the North Island. No group is entirely absent from the South Island. Occurrences outside the two main islands but still within the New Zealand Botanical Region are indicated as follows.
Distribution beyond the New Zealand region, not shown in the diagrams, involves only two of these groups of Hebe. In South America there are two species of ‘Apertae’ (indistinguishable from South Island plants of H. salicifolia and H. elliptica) and one of these (H. elliptica) extends to the Falkland Islands. On Rapa Island, west of Pitcairn, is found an endemic species belonging to ‘Occlusae’; it resembles Chatham Island species closely, but shows less affinity with any species of New Zealand proper.
The second set of diagrams (Fig.2) depicts leaf bases. In Hebe leaves are always in opposite pairs, and it is characteristic of the genus that the leaf bud at the tip of the shoot is large; the two outermost leaves of the bud are almost full grown before they break apart (well illustrated by photographs in Wardle, N.Z. Jl Bot. 1, 1963: 34). The closed leaf bud gives a good clue to the group to which a plant belongs. First identify the midribs and the margins
Figure 3 shows diagrammatically some features of capsules of Hebe, Pygmea and Parahebe. The capsule valves are shown in outline transverse section; the septum between the two loculi is placed vertically and the floral bract and axis are shown in the conventional way by a triangle at the left and a dot at the right
Hebe from Veronica, a short septum being characteristic of the latter genus.
Figure 4 illustrates inflorescence types. The diagrams also show the three leaf forms that result from the types of buds shown in figure 2, i.e. petiolate (as in ‘Buxifoliatae’), sessile (as in ‘Occlusae’) and connate (as in ‘Flagriformes’ and ‘Connatae’).
A first point to notice about an inflorescence is whether it is strictly terminal or wholly lateral. In the latter case the apical leafy shoot (shown by an arrowhead) is available to carry on vegetative growth above the inflorescence. If the inflorescence is terminal, new leafy shoots can come only from axillary vegetative buds below the inflorescence. When flowers are clustered near the end of a twig, some care is needed to be sure whether or not there is a leafy tip hidden between them.
The inflorescence is racemose though occasionally the pedicels are so short that it appears to be a spike. The least elaborate is that seen in ‘Buxifoliatae’ and ‘Flagriformes’ where the bracts are opposite and almost as large as the leaves, so that the flowers can almost be regarded as solitary and axillary. In species or whole groups where the bracts are smaller and less leaf-like they show an increasing tendency to lose their opposite arrangement — a tendency
Pygmea and in some species of Parahebe are each associated with two, usually opposite, bracts and can probably be regarded as representing much reduced racemes. Such a very simplified inflorescence has not been seen in Hebe.
On the basis of these and other similarities and contrasts, possible relationships between groups are suggested by connecting lines in the diagrams, and in part also by the relative positions of the insets. If the original ‘protohebe’ is imagined as something not unlike Hebe odora in ‘Buxifoliatae’, the present representatives could be derived by gradual specialisation of the reproductive shoots. As bracts are more and more different from leaves in size, texture, and arrangement, the inflorescence is correspondingly more isolated into lateral racemes or terminal panicles. Minor changes can be induced experimentally. It has been noted that in rather pampered garden plants of
For more than a hundred years artificial hybrids have been made between species belonging to widely different groups of Hebe but results have rarely been precisely recorded. A systematic programme of crossing might rather quickly indicate whether appropriate weight has been given to the characters here used to group species together and to differentiate between the groups. Obviously there is still much to be learned from looking hard at Hebe.
The GenusFicus, best known in temperate regions by the species providing the fig, comprises about 800 species which are widely distributed in the warmer parts of the world. The genus is conspicuous in lowland tropical rain forests by virtue of its wide range of growth habit, many species coming into the category known as strangling epiphytes, others being root climbing lianes and others again ground rooting trees or shrubs. The latter produce aerial roots very freely and in this respect the extreme case is the ‘banyan’. and other species of similar habit, whose branches, supported by aerial roots, may spread out over acres of ground.
From my knowledge of the New Zealand lowland forest and from visits to New Caledonia and Fiji it has occurred to me that Metrosideros, with its several strangling epiphyte species and root climbing lianes, as well as ground rooting trees and shrubs, provides a remarkable if less exuberant ecological parallel to Ficus.
Metrosideros is a genus of the Myrtaceae, notable for the brightly coloured flowers of some species, which ranges from New Zealand to Malaya and Hawaii with an isolated species in South Africa. Thus its geographic range is much less than that of Ficus, and it also encompasses fewer species, perhaps 50, although this figure would be increased if the related genus Mearnsia were to be united with it. In New Zealand Metrosideros is largely restricted to the lowlands, and in the tropics largely to montane rain forests. Metrosideros is thus suited to a lower temperature regime than Ficus, a point which is emphasised by the presence of protective bud scales in a number of the species.
Of the eleven New Zealand species one — Metrosideros robusta — is a strangling epiphyte. This species is common and conspicuous in the rain forests from the far north to the north of the South Island. It establishes on a wide variety of host species, but is most common on the emergent conifers and, by virtue of this, itself becomes a tall emergent. The details of its growth habit and in particular whether or not it kills its host have been the subject of argument. Kirk (1872) claimed that the roots of the Metrosideros often enclose the host trunk and kill its living tissues by compression.
Metrosideros has no significant effect on the host. It does not establish until the host is mature, so by the time it has achieved independance the host might well have died of old age (Zotov 1948).
It certainly seems clear that the roots of Metrosideros robusta do not completely enclose the host trunk as is the case with many of the ‘strangling fig’ species. In the latter the young descending roots form a complete network around the host trunk, but in Metrosideros robusta the one to several vertically aligned roots are usually disposed to one side of the host trunk, except near the ground where they often enclose it with branch roots. In the simplest case there is a single main root, which is attached to the host in the early stages by a number of slender, horizontal ‘girdling’ roots (Fig 1). The main root often branches near the ground to form a tripod arrangement. (Fig. 1; Fig. 2, left). Thus the growth habit of Metrosideros robusta is very similar to that of the New Zealand shrub epiphyte Griselinia lucida (Dawson 1966), although the latter does not appear to become self supporting.
The habits of Metrosideros species outside New Zealand are not so well known, but at least one species in New Caledonia is a strangling epiphyte and has a very similar growth habit to Metrosideros robusta (Fig. 3, left). The sole Fijian species — M. collina var. vitiensis — is also frequently a strangling epiphyte in montane forests (Fig, 3, right). This species is notable in that it also plays a quite different role as a terrestrial shrubby pioneer following fire and in this regard seems to provide a parallel with Leptospermum in New Zealand. Metrosideros robusta in New Zealand may also occur terrestrially, but not so commonly as the Fijian species. Possibly two species of Metrosideros in the New Hebrides are strangling epiphytes and there may be others elsewhere. In general, in Fiji and New Caledonia at least, Ficus ‘stranglers’ give way to Metrosideros ‘stranglers’ at higher elevations, but there is considerable overlap and in both places I have noted instances where Ficus and Metrosideros shared the same host.
Three of the terrestrial tree species of Metrosideros — M. excelsa (pohutukawa), M. umbellata (southern rata) and M. kermadecensis — resemble many of the terrestrial Ficus species in their ability to produce aerial roots. Some individuals of Metrosideros excelsa in particular may produce an abundance of pendant roots from their spreading branches (Fig. 4). However these roots do not usually
Probably none of the New Zealand species can be regarded as true shrubs. M. perforata and to a lesser extent, M. diffusa assume a shrubby habit when they establish in open sites, but they are typically forest lianes. It is interesting to note that some tropical lianes also grow and flower as shrubs in similar circumstances (Richards, 1952 p. 105). M. excelsa, M. kermadecensis, M. umbellata and M. parkinsonii can flower when they are quite small, but they all achieve the dimensions of trees, even though the last is only a small tree and in some localities may persist for a long time at the shrub stage.
In New Caledonia on the other hand there are several species, mostly undescribed, which are true shrubs. These commonly grow along stream margins, or even on rocks in the middle of streams, and have a four ranked leaf arrangement which is very reminiscent of Hebe in New Zealand. I am not aware of shrub species of Metrosideros anywhere else.
Six of the New Zealand species of Metrosideros — M. fulgens, M. perforata, M. diffusa, M. colensoi, M. carminea and M. albiflora — are root-climbing forest lianes very similar in habit to the climbing species of Ficus. Seedlings of these Metrosideros species establish on the ground and their slender stems attach themselves by adventitious roots to tree trunks. The leaves at this stage are in most of the species flattened against the host trunk to form a mosaic (Fig. 5, inset).
Eventually the ascending stems reach the tree crown seventy or more feet above ground level and in the full light send out many more or less horizontal branches which bear the flowers.
The climbing stems lower down lose their attaching roots and separate from the host trunk. The species vary in the maximum size attained by these stems but those of Metrosideros fulgens and
I am not aware of liane species of Metrosideros elsewhere than New Zealand although some species of Mearnsia in New Guinea are described as root climbers (White 1942).
Although a few species of Ficus are cultivated quite successfully out-of-doors in the warmer parts of New Zealand there are no native species, nor is there any evidence that the genus occurred in New Zealand in former geological times. Probably in the past as now Metrosideros played a role in the New Zealand rain forests parallel to that of Ficus in tropical rain forests. Was New Zealand the centre of origin, or at least diversification of Metrosideros with later migration into montane tropical areas? This seems a likely possibility, but the question cannot really be seriously considered until the genus and others related to it have been studied in greater detail, particularly in the tropical areas.
The Following Description of the male has been based mainly on Owen (1841), who covered comparative anatomy more widely than later authors. Other references are Huggins and Potter (1959) and Pavaux (1962).
The vasa deferentia pass the ureters laterally and posteriorly to join the anterior end of the urethra. There are no vesiculae seminales, but the epididymes are large. The prostate is larger than that of placentals. It forms a single body, the urethral bulb, which is enclosed in a sheath containing mainly transverse muscle fibres. Two or three parts, arranged serially, can be distinguished macroscopically by structure or by colour. Its secretions discharge into the urethra through many pores. There are two or three pairs of Cowper's glands which communicate with the urethra by a common duct on each side.
The corpora spongiosa originate in two distinct muscular bulbs (placentals have one). The bulbs of the corpora cavernosa are enclosed by the erector penis muscles which are attached to the pubes (Fig. 3). Retractor penis muscles arise near the middle of the sacrum. A levator penis muscle is found in some species as a branch arising on each side from the fascia of the crus penis and converging ventrally to form a single tendon which joins the penis.
The glans may be bifurcated or single. In the former case each lobe is transversed by a seminal groove or an enclosed duct (Biggers 1966).
The main parts of the generalised female reproductive system are shown in Fig. 4. Names of the organs are not completely standardised, but those chosen conform to known functions. Like the prostate of the male, the vagina is large.
There are always two separate uteri. The vagina has three parts, a median sac and two lateral canals. The vaginal sac surrounds the ora uteri. It is at first divided dorso-ventrally by a septum which may break down during late juvenile life or at any later stage. Sometimes the septum may remain permanently. Variation occurs both among and within species. The sac communicates during late pregnancy with the anterior end of the urogenital sinus by a median
et al. 1964). The median canal has no intrinsic muscles, and the lateral canals, which are muscular throughout most of their length, also lack these within the urogenital strand. However, circular muscles, anterior to the urogenital sinus, surround all the vaginal canals in the periphery of the strand. The ureters pass between the lateral canals and the median sac and discharge through two papillae within the neck of the bladder. The urethra enters the urogenital sinus ventrally.
The median canal is usually a simple unlined slit which forms late in pregnancy and closes soon after parturition. In some species, mostly the Macropodidae, this birth canal may remain permanently open and it is then lined with epithelium. Infrequently, a temporary median canal may be lined similarly (Kean et al. 1964).
In many species the lateral canals are occluded posteriorly by epithelial proliferation Short term closure of the lateral canals does not necessarily depend on cellular growth. Valves are present in some Macropodidae (Owen 1841). In Trichosurus the canals are usually blocked after copulation by an acellular plug derived mainly from the prostate (Kean et al. 1964), and the presence of an additional ring of circular muscles at the posterior ends of the lateral canals in this species suggests alternative muscular occlusion. Vaginal occlusion occurs in many placentals (Asdell 1946; Kean 1961).
The pattern of the marsupial vagina is varied by differences in expansions either of the lateral canals anteriorly, or of the median sac. Lateral canals are long in the Caenolestidae and the whole vagina tends to be long in the Australian marsupials.
In most marsupials placentation is of the yolk sac type (chorio-vitelline). In early stages there is a bilaminar omphalopleure. Mesoderm expands progressively to varying degrees between the two layers. In some marsupials at least (Flynn 1938-39; Sharman 1965a) close apposition of the placenta with the uterine wall during much of the gestation period is prevented by an apparently attenuated shell membrane. A syncytium derived from foetal and maternal cells, or sometimes possibly from maternal cells only, occurs in some species. An allantoic placenta is found in Perameles and Thylacis (Hill 1899) and in Phascolomis and Phascolarctos (Amoroso 1955), but detail for the last two species is not available.
Classification of marsupial placentation (discussed by Sharman 1959b, 1965) is subject to difficulties because type of placentation is not evidently associated either with neonatal size or with evolutionary status of the species, but the situation in placentals is equally unsatisfactory. Young (1957) pointed out that in the Placentalia the evidently primitive diffuse and epithelio-chorial type of placentation occurs among the most specialised mammalian orders. If types of placentation were used as a basis for taxonomy, lemurs should be combined with ungulates (because their placentation is epithelio-chorial) and the hyraxes with the insectivores, primates and other haemo-chorial types.
In view of the disconformities of mammalian placentation, consideration can be given to placentation in viviparous lizards and snakes, which seem to approach the probable condition of the ancestors of the present day mammals when they were changing from oviparity to viviparity (Weekes 1935). Despite independent origins among different reptilian genera, placentation is very uniform. Three evolutionary steps were recognised in modern lizards. In the first of these, corpora lutea developed and allowed the eggs to be retained in the uterus until completion of embryonic development. In the second,
Early establishment of the yolk-sac did not affect placental development. Where the yolk-sac intervened between the allantois and the uterus the intervening yolk-sac tissues — glandular epithelium with mesoderm and endoderm — soon degenerated, giving way to allantoic placentation (Weekes 1935).
Weekes (1935) suggested that in reptiles the first function of the yolk-sac placenta was the absorption of fluid, to facilitate yolk conversion following reduction of albumen in uterine eggs. Since blood vessels were not increased in the absorptive yolk-sac, the vascular allantois came to form the main placenta in the third (nutritive) stage described. But among mammals, in a fourth step, vascularisation of the placentary surface of the yolk-sac has occurred, either by the distal half of the placenta developing blood vessels, as in most placentals, or by this half becoming eroded, so permitting the vascular inner surface of the yolk-sac to be applied to the maternal endometrium, as in the inverted yolk-sac placenta of rodents (Young 1957).
Although the respiratory allantois, by its contact with the inner surface of the shell membrane and by its external vascularisation, was pre-adapted for placentation, the yolk-sac placenta became equally efficient after it had developed external vascularisation. Reduction in size of the marsupial embryo would have reduced placentary requirements and then the yolk-sac placenta, because of its earlier development in ontogeny, would have been retained rather than the allantoic one which, in most genera, has reverted to its initial primitive function as a receptacle for urinary waste.
Marsupials, with the possible exception of Dasyurus (see Hill and O'Donoghue 1913) are polyoestrous. Ovulation is spontaneous.
Lactation at the end of the oestrous cycle can be induced in virgin females by the stimulus of suckling more readily in marsupials than in placentals, indicating little physiological difference between the oestrous cycle and pregnancy in marsupials in regard to lactation (Sharman 1962). It has been suggested (Sharman 1955) that the corpus luteum formed when post-partum oestrus occurs has a lactational function, but most marsupials do not have such ‘corpora lutea of lactation’. In general, it seems that the minor difference in lactation between marsupials and placentals arises from the short gestation typical of the former group and results in early lobule and alveolar development of the mammary glands.
In view of lactational similarities, Sharman (1965a) suggested that gestation in marsupials required only the hormones present during the oestrous cycle. This is doubtful since in pregnancy the post-ovulatory Graafiian follicles disappear insted of continuing on to ovulation, and the reduced vagina does not undergo pro-oestrous expansion (except in species specialised for post-part oestrus).
In Didelphis failure of pregnancy ofter ovariectomy (Hartman 1925) suggested that marsupial embroyos did not produce, from the placenta or elsewhere, hormones which could augment those of the corpus luteum, but Tyndale-Biscoe (1963b) found in Setonix that ovariectomy after the sixth day from coition did not prevent full development and retention of the embryo although parturition was unsuccessful. These results await further work for adequate explanation.
Post-partum oestrus occurs in many of the Macropodidae (Sharman 1955, 1963; Hughes 1962a) and probably in the phalanger Cercaertes concinnus (Bowley 1939). In such macropods parturition is immediately followed by another pregnancy in which intra-uterine development is suspended at an early blastocyst stage, but growth is resumed if suckling young leave the pouch. Decline of reproductivity to anoestrus may cause the blastocyst to be resorbed (Sharman 1955) or it may simply prolong suspension of growth until conditions are favourable for reproduction (Berger 1966).
Prolonged gestation in macropods is a favourable specialisation for unpredictable climatic aridity (Ealey 1963; Newsome 1965). During drought, females of Megaleia lose youch young before they are forced into anoestrus, and presence of a uterine blastocyst increases the number of successive young which can be produced in adverse conditions. Such young require a minimum of maintenance because of their minute initial size.
The reason for prolonged gestation in Cercaertes is unknown, and it is uncertain whether it conforms to the macropod pattern.
Viable sperm are not retained in the marsupial vagina, but Hill and O'Donoghue (1913) found sperm closely packed in parallel formation in gland lamina of the Fallopian tubes of Dasyurus. It is not known where spermatozoa are held similarly in other marsupials, but the method of sperm storage described by Hill and O'Donoghue was illustrated by Fox (1956) in photomicrographs of gland lumina of oviducts anterior to the uteri of snakes. (Captive lizards, snakes and turtles, in the absence of parturition, could retain viable sperm for several years.)
In the early pouch young, marsupials are clearly differentiated from placentals, in both sexes, by the Wolffian ducts, which pass the ureters laterally instead of mesially.
The Wolffian ducts develop early in both male and female embryos. The ureters arise later, in apposition to the metanephric blastema, as dorsal or lateral buds (Fig. 5)t on the Wolffian ducts (Buchanan and Fraser 1918; Hill and Hill 1955) and become extended as embryonic growth carries the kidneys to relatively anterior positions. Müllerian ducts form subsequently as infundibula which extended posteriorly from the mesonephric region. Pearson (1947) pointed out that the Müllerian ducts reach the urogenital sinus by following along the route already established by the Wolffian ducts, as described by Baxter (1935), and he suggested that their positions in marsupials and placentals was a consequence of the positions of the Wolffian ducts. His hypothesis was supported by re-examination of material used by Buchanan and Fraser, and by further reseach, which showed than in many, if not all, marsupials the Wolffian ducts contributed posteriorly to the actual formation of the lateral canals (de Bavay and Pearson 1949).
Since there is no apparent reason for differences in the placement of marsupial and placental Wolffian ducts in males, the positions of the Wolffian ducts seem to have been determined by the respective female requirements.
Since the mentanephric blastema are invariably situated mesially differences in positions of the Wolffian ducts are sufficient explanation for dorsal, mesial or lateral orientation of the buds on these ducts. The orientation has no apparent significance because allometric growth of the cloaca and the bladder primordium determines the final positions of the ureters. These ducts, in fact, occupy identical positions in marsupials and placentals — between kidneys and
The Müllerian ducts of the marsupial are initially straight (Fig. 6) and would allow unrestricted passage to small marsupial young during parturition, but the Müllerian ducts, together with the adjacent Wolffian ducts, subsequently become looped laterally and anteriorly (Fig. 7). This looping increases the capacity of the Müllerian ducts
A further change of form occurs in the female: the Müllerian ducts approximate mesially and together extend a short distance posteriorly, but they do not communicate with the urogenital sinus until the approach of sexual maturity or parturition. Posterior mesenchyme of the uterine embryo and the early pouch young becomes differentiated into the muscles of the lateral canals and the sac, and into connective tissue and muscle fibres of the urogenital strand. The ureters also move laterally and leave unoccupied the connective tissue which accommodates the later extension of the median vaginal canal or birth passage.
Marsupials utilise food throughout the wide range eaten by the placentals, by parallel adaptions. In general, the caecum and the gall bladder are large (Owen 1841) and the gut shows the
Dromicops (Didelphidae) and in the nectar-feeding Tarsipes (Phalangeridae) (Hill and Rewell 1954). In Vombatus the caecum is vestigial, but the anterior end of the colon is enlarged and caecum-like, suggesting adaptation for readily digested foods, followed subsequently by a return to fibrous material and re-adaptation for this (Lonnberg 1902; Hill and Rewell 1954).
In the Phalangeridae the gut and caecum become relatively longer with increase in body size and decrease of insects in the diet. In the anomalous Phascolarctos, which has affinites with both the Phalangeridae and Vombatidae, the caecum is very large and presumably digests cellulose efficiently. Pseudocheirus, with a sacculated caecum having two very strong taenia, is probably also efficient. Trichosurus has a relatively simple caecum divided partially by constrictions into four pockets. It is still large although relative to body size only half the length found in Pseudocheirus. It appears to digest parenchymatous leaf substance but not the more fibrous tissues (Lonnberg 1902). Honigmann (1941) reported high values for absorption of crude fibre by Trichosurus when tested on banana and carrot, but such material would conform only to the parenchymatous material mentioned by Lonnberg, and over-all digestion of fibre found in more natural foods is probably relatively poor since this species in the wild state avoids fibrous material (unpublished data). Petaurus, which supplements vegetable material with insects, screens all large particles from entering the caecum (Lonnberg 1902).
In Trichosurus it is likely that rate of caecal digestion is slow, resulting in a high proportion of food material being passed directly from the small intestine to the colon because of closure of the ileocaecal valve. D. Gilmore (personal communication) found that food was eliminated in from eight to more than 120 hours, with 80% passage in thirty-six to ninety-two hours.
In the allied Macropodidae specialisation for grazing has resulted in more complicated structure of the stomach and the adoption of ruminant-type microbial digestion. Moir et al. (1956), for Setonix, described partial division of the stomach into four parts. There is a sacculated fore-stomach (the putative rumen), a non-sacculated region to which a well defined groove leads from the oesophagus, a highly acidic region analogous to the ruminant abomasum, and a smaller fourth region of unknown function. Calaby (1958) found that Setonix digested fibre considerably less efficiently than did sheep and cattle. In protein assimilation the animals were much the same, but food residues were passed more quickly by Setonix. Foot and Romberg (1965) found that Megaleia rufa ate relatively more oat
Megaleia assimilated less crude fibre but showed a higher total intake of nitrogen — apparently associated with a more rapid passage of food through the alimentary tract (as also in Setonix). This advantage appeared to be offset somewhat by a higher rate of nitrogen excretion, but Megaleia might have the lower basal metabolic rate since sheep were less successful in maintaining body weight on a diet of straw.
The utilisation of low protein foods was demonstrated in Western Australia where Macropus robustus lived in the Triodia covered hills while Ma. canguru was found on the plains where more nutritious grasses were available. Droughts and grazing by sheep resulted in Triodia, which has a low nitrogen content, becoming dominant on the plains, and this change permitted Ma. robustus to displace both Ma. canguru and sheep (Main et al. 1959). In competition with sheep, Ma. canguru would probably have been the more successful since, from preference or unselective feeding, it takes a higher proportion of coarse grass (Kirkpatrick 1965), but its minimum requirements seem to be greater than those of Ma. robustus.
Ma. robustus produces a urine of high concentration, and although the animal drank by preference, a proportion of the population was capable of surviving for long periods without free water if shade was available (Ealey et al. 1965). This faculty was an obvious advantage in an arid climate, but Ealey (1963b) suggested that it was also associated with an adaption for a diet low in proteins. Schmidt-Neilsen (1964), for the camel, confirmed earlier reports of high urea secretion falling when water supplies were limited and he found it as low as 1/50 of normal level. He suggested that this change resulted from the passing of urea to the stomach where the microflora re-converted it to protein. Similarly, in Macropus robustus increase in drinking water resulted in increase of nitrogen in the urine Ealey 1963a).
The macropods appear to differ from domesticated ruminants, not so much in digestive efficiency, as in adaption to the coarse forage of arid land.
Adequate treatment of brain structure and psychology would extend far beyond the scope of these notes. Literature on structure is extensive but mostly concerned with fine detail. Comparative morphology was covered by Owen (1841).
The brains of many marsupials are small, but some brains are large by equivalent placental standards. The most notable feature is the absence of a corpus callosum, the main inter-cerebral
Such apparent commissural deficiency fostered a belief of marsupial primitiveness, but recent work (Sperry 1964; Gazzaniga 1965; Gazzaniga et al. 1965) showed that in placentals severance of the corpus callosum had little effect on intelligence. The commissure was associated with cerebral dominance and lateralisation, which is most highly developed in man. These terms denote specialisation of functions in one or other of the hemispheres — as speech on the left and spatial cognition on the rigtht. Visual perception and some tactile sensations are lateralised, but some stimuli are conveyed to both hemispheres, and there are traces of inter-cerebral communication even when all commissures have been cut. Accordingly, there is some need for lateral dominance, which is illustrated in a simple form by right handedness.
Among species, intelligence is largely proportional to cerebral size, and some marsupials could not be expected to score highly, but even small-brained marsupials have been greatly under rated because of comparisons with placentals, made without regard to marsupial specialisations (Hediger 1958). It now seems that the assumed deficiency of commissures in marsupials may not be primitive, but that it indicates bilateral organisation to a degree greater than found in placentals.
History of evolution is always tentative. Deductions are subject to the limitations of available evidence.
The early mammals are not shown to have differed greatly from reptiles in brain structure. Their main advance was endothermy, with which was associated insulating hair. A sweat gland system assisted in maintaining condition in hair and skin, and produced odours of social importance. Many mammals were too small to sustain the water loss required for evaporative cooling by sweat. Milk evolved from sweat glands and is itself evidence of pre-established socialisation in care of young.
Viviparity was almost certainly present in some reptiles, but it has not been shown for early mammals. It would have been advantageous to heliothermic animals, rather than to endothermic ones. In mammals it probably arose late, as a consequence of socialisation which afforded maternal protection to free young.
Such young would have been relatively large, preceded in evolution by heavily yolked eggs which produced the yolk-sac and respiratory allantois that were to form the placenta. Efficiency of a simple placenta depends on the relationship of placental surface to bulk of the embryo, decreasing as an embryo increases, and the
A single functional oviduct is present in birds and monotremes. Probably analogous development of a median vagina occurred in oviparous therians during some 50 million years preceding separation of marsupials and placentals. Eggs were hard shelled according to the vestigial evidence of a caruncle on the marsupial embryo.
In many recent marsupials the lateral vaginal canals receive a large volume of semen which cannot be accommodated in the median passage. In the other marsupials, semen is transported to the median sac, probably as a secondary development which evolved simultaneously with anoestrous closure of a narrow median canal. The width of this canal in living marsupials is commensurate with neonatal size, suggesting reduction associated with that of the neonatus. But a small therian would have had a relatively large birth passage, able to accommodate semen.
Common therian inheritance would explain the presence of a median vaginal canal in both placentals and marsupials. This origin has not always been accepted. Sharman (1959a, 1965b) agreed that the median canal is not needed by recent marsupials, but stated that the ‘more posterior pseudo-vaginal canal is a new, purely marsupial, development’. No reason or phylogenetic antecedent was given for such new development, and similarity of lateral canals, median canal, and placental vagina was not explained. These canals of three types are all derived directly or indirectly from Müllerian ducts and the cloaca, with varying participation of the Wolffian ducts (Buchanan and Fraser 1918; Baxter 1935: Kean et al. 1964). Admittedly the posterior median canal originates anteriorly from the vaginal sac after it has become differentiated from the Müllerian ducts, and the canal is completed late, after sexual maturity, but late completion is somewhat misleading. Organisation for the canal is early, as shown by the lateral deflection of the lateral canal primordia and the ureters, which leave a clear path for the canal. The lateral canals require little explanation; phyletic antecedence is unnecessary since they are functional.
Sharman (1965b) erroneously cited Kean (1961) as claiming that marsupials differ from all other amniote animals in the course followed by their Wolffian and Müllerian ducts, and the statement (Sharman 1965b) ‘The ureters take a path between the Wolffian ducts in monotremes, reptiles and marsupials so this is the ancestral condition’ was also incorrect. In reptiles and monotremes the ureters lie dorsally and end in the wall of the uregenital sinus. Marsupials and placentals differ from all other amniotes in that their ureters cross from the dorsal position to a ventral one where they join the neck to the bladder. Since this feature is present in all therians, it must have been an early
Differences between marsupials and placentals in the courses followed by the Wolffian ducts are readily explainable. From an early therian vagina with centrally united Müllerian ducts, marsupials would have required only to develop lateral canals. Their lateral positioning of the Wolffian ducts would then have been a simple case of neoteny—facilitation of female embryonic development — as already explained, of no selective consequence to the fully formed adult.
The parallel evolution of allantoic placentation in many reptiles, the absence of similar yolk sac placentation, and the pre-adaption of the externally vascularised allantois, indicates that the occasional allantoic placentation among marsupials is vestigial and not rudimentary.
Reduction of neonatal size seems to be an inescapable conclusion. The 0.012g neonatus of Dasyurus would be improbably small for a primitive non-lactating therian, and the young to adult weight ratios of marsupials are clearly separable from placental ones. Sharman (1965b) postulated a present evolutionary trend for increasing neonatal size, but adequate published data are available for only three species other than macropod ones, and although neonatal size increases in macropods, relative neonatal size tends to decrease.
Reproductive hormones of marsupial females show considerable resemblance of those of placentals, with due allowance made for shortness of gestation. Oestrous cycles usually exceed gestation in length, except in Macropus canguru (Poole and Pilton 1964) and in Protemnodon bicolor (Sharman et al. 1966). Asynchrony of oestrous and gestational periods has been recorded for Potorous (Flynn 1922).
Prolonged gestation occurs in many macropods as an adaption to arid country with unpredictable rainfall. Embryonic development which follows postpartum oestrus is totally suspended at an early blastocyst stage but is resumed if youch young are lost while conditions are suitable for reproduction.
Regulation of body temperature is efficient. The main difference from placentals is that marsupials tend to have more definite active and resting temperatures, but there is no clear distinction between the two groups in this respect.
Digestive specialisation of marsupials are comparable to those of placentals. Grazing macropods have ruminant-like macrobial digestion, and they show a high degree of adaption to coarse grasses.
Many marsupials have small brains, but others are fully equivalent to placentals. The corpus callosum is absent. This feature
Marsupials have been widely adaptive. Much of their evolution has paralleled that of placentals, and in a number of respects they demonstrate alternative efficient methods methods of attaining the same ends.
I am indebted to Dr W. D. L. Ride and Mr J. M. de Bavay for discussion and assistance, to Dr R. A. Barbour for advice on musclature, to Mr and Mrs F. Farrar for German translation, and to Miss J.
Marsupials mentioned
Didelphis virginiana — Virginia opossum
Marmosa spp. — Murine opossums
Metachirus nudicaudata — Brown opossum
Dromicops sp. — Opossum
Antechinus flavipes — Yellow-footed marsupial mouse
Dasyurus (viverrinus) = quoll — Native cat
Dasycercus cristicaudata — Crest-tailed marsupial mouse
Myrmecobius fasciatus — Banded anteater
Phascogale tapoatafu — Black-tailed phascogale
Satanellus hallucatus — Little native cat
Sarcophilis harrisii — Tasmanian devil
Sminthopsis crassicaudata — Fat-tailed marsupial mouse
(Isoodon) = Thylacis macrourus — Brindled bandicoot
Macrotis lagotis — Rabbit bandicoot
Perameles nasuta — Long-nosed bandicoot
Cercaertus (= Cercartetus = Dromica) spp. — Pigmy opossums
Eudromicia spp. — Pigmy opossums
Pseudocheirus peregrinus — Ring-tailed opossum
Trichosurus vulpecula — Brush-tailed opossum. (Widespread in N.Z.)
Phascolarctos cinereus — Koala
Phascolomis mitchellii = Vombatus hirsutus — Common wombat
Hypsiprymnodon (= Hypsiprymnus) moschata — Musky rat kangaroo
Macropus canguru (= major) — Grey kangaroo
Macropus robustus — Euro
Megaleia rufa (= Macropus rufus) — Red kangaroo
Petrogale penicillata — Bushtailed rock wallaby. (Kawau and Rangitoto Islands, N.Z.)
Potorous tridactylus — Rat kangaroo
Protemnodon rufogrisea (= Macropus ruficollis) — Brush wallaby. (South Canterbury, N.Z.)
P. bicolor — Swamp wallaby. (Kawau Island, N.Z.)
P. eugenii — Tammar. (Kawau Island and Rotorua, N.Z.)
Omitted from P. parmaIntroduction because this species was first identified as established in New Zealand by Dr Protemnodon dorsalis was previously rare on Kawau Island and is now likely to be extinct.
Setonix brachyurus — Quokka