Vegetation of Tern and Gannet Colonies in Northern New Zealand with a Comparative Note on Colonies in the Bass Strait, Tasmania

Transactions and Proceedings of the Royal Society of New Zealand, Volume 88, 1960-61, Page 211

Vegetation of Tern and Gannet Colonies in Northern New Zealand with a Comparative Note on Colonies in the Bass Strait, Tasmania By Mary E Gillham, Botany School, University of Melbourne* This work was carried out during the months before and after an eight-month period of lecturing at Massey College, University of New Zealand. [Received by the Editor, September 30, 1959.] Contents Summary Introduction I. White Fronted Tern Colonies. (1) General Observations. (2) Colonies in the Hauraki Gulf. (3) White Island, Bay of Plenty. (4) Summary of tern vegetation and comparison with that of other areas. II. “Natural” Gannet Colonies. (1) General observations. (2) Colonies in the Hauraki Gulf. (3) White Island colonies. (4) Black Reef colony, Hawke's Bay. (5) Summary of “natural” gannet vegetation and comparison with that in Bass Strait. III. “Exotic” Gannet Colony of Cape Kidnappers. (1) History of the gannet occupation. (2) Zonation of vegetation adjacent to gannets. (3) Floristic composition of vegetation adjacent to gannet colonies. Summary Deposition of guano and trampling by terns and gannets has a suppressive effect on vegetation leading to the elimination of much of the less halophytic, indigenous flora. Guano, unlike salt spray, is deposited only seasonally, and as the soil solution becomes less concentrated after the breeding season there is an influx of winter annuals. Most of these are non-halophytic, and this influx emphasises the relative importance of guano and sea spray in suppressing the plants, as the latter is likely to be more abundant (though concentration due to evaporation less) when the annuals are present. Almost all the annuals are alien species, even on remote stacks, and in Cape Kidnappers, a mainland gannetry, only 8 per cent of the species within 5 m. of the nesting area are native. Plants surviving the sulphurous volcanic fumes on White Island appear to be also eminently suited to withstand wind-blown sea salt and guano. Gannets are much more destructive of vegetation than are terns, and the most resistant perennials are Disphynia australe and Coprosma repens. Parallels are drawn with tern and gannet colonies in the Bass Strait, Australia. Introduction TErns and gannets are gregarious and the vegetation of their nesting and roosting colonies is intimately affected by their trampling and manuring. Those plants which survive must be morphologically and physiologically adapted to intense seasonal disturbance and alterations in the composition of the soil solution. Such.

species are relatively few and the same ones occur in widely separated colonies and at varying heights above sea level. In order to survive the summer nesting season a plant must have the rigidity or lowly habit which suffer a minimum of mechanical damage and must be tolerant of high concentrations of guano in the soil. These two features are best shown by halophytes of windswept situations and salty soils, so “bird vegetation” resembles “sea spray vegetation”, even when both modifying factors are not operating simultaneously, as is often the case. Another group of plants represented in the bird colonies is the annual or ephemeral vegetation which invades seasonally when the birds are not in residence and dies out progressively as the nesting season advances (Gillham, 1956). The plant communities of tern and gannet colonies fall into two categories—“natural” and “exotic”. The first includes all the tern colonies, and the gannet colonies of Hauraki Gulf, Bay of Plenty and Black Reef, near Cape Kidnappers. It may be subdivided into those areas which would normally bear a herbaceous rock vegetation with a high proportion of succulents and those which would normally bear maritime scrub or bush. The second vegetation category is exemplified by the exotic grassland community bordering the two main portions of the Cape Kidnappers gannet colony, near Napier. I. WHITE FRONTED TERN COLONIES 1. General Observations Sterna striata (Gmelin, 1789), the white fronted tern, nests in the same type of coastal rock habitat as does Larus novae-hollandiae Scopulinus (Forster, 1844), the red-billed or mackerel gull of New Zealand and the silver gull of Australia, and also on shingly river beds. No colonies on river shingle were visited, and of the coastal colonies seen most were on low sea-girt stacks. On the lowest stacks there was often no vegetation, due to the combination of sea spray and bird activity; elsewhere, from about 3 m above sea level in sheltered waters to well over 30 m. on White Island, the vegetation was usually dominated by Disphyma australe (=Mesembryanthemum). Terns appear to have a less deleterious effect on the indigenous plants of the breeding area than have gulls, two of the contributory factors being as follows— (a) Often little or no nest material is used, so portions of plants adjacent to the breeding site are not severed, and those plants actually on it are not so completely smothered by material brought from elsewhere. Considerably less than half the birds seen had made any attempt to build a nest, these using dead Disphyma stems from the vicinity, occasionally with a little dead grass incorporated. In the majority of instances the eggs had been laid directly on the underlying rock or plant cover, and many had rolled from these precarious situations into narrow crevices or under bushes where they could not possibly be incubated. None of the eggs had hatched when the islands were visited between late November and early December, 1957, and the underlying Disphyma was often little harmed. (b) Terns feed at sea and have no opportunity to bring alien weed seeds back to the nesting area as have gulls with their more omnivorous feeding habits. The non-native element in the flora is much more noticeable in gull colonies, but the partially bare, periodically disturbed and well manured soil of tern colonies other than those on dense carpets of Disphyma forms a suitable habitat for short-lived ruderals should disseminules arrive. Phalacrocorax punctatus punctatus (Sparrman, 1786), grey or spotted shags may occupy areas alongside the terns but return to these throughout the year and like the gulls, do more damage to the vegetation. On David Rocks in the Noises.

Group, both species penetrated further back beneath the coastal scrub than was usual. In the tern area the undergrowth beneath the shrubs was relatively unharmed; in the shag area it was almost non-existent, and the shrubs had become denuded of their lower branches to give a “leggy” scrub emerging from soil white with guano and puddled hard by trampling. 2 Colonies in the Hauraki Gulf The densest population of terns was found on Crusoe Island, between the islands of Motuihe and Waiheke, in the Hauraki Gulf. The birds were breeding on one tidal islet off the south of the main island and three off the north. In each case the average distance between the single or paired eggs was only 30 cm., the sitting birds orientated in the same direction, head to wind, with little room between. Some were in well-worn hollows which appeared to have been occupied in previous years, others were in “virgin” habitats on fresh vegetation. On the south stack the dominant Disphyma australe was dead over a considerable area of the summit, decay having proceeded too far for this to be a result of damage during the current nesting season only. It had succumbed both on the ridges, where soil depth was negligible and strong winds would hasten desiccation, and also in the hollows to which the accumulated manure drained. This summit area would probably be used as a standing ground after the young had left the nests, to a greater extent than would the sloping sides where the Disphyma was green and healthy, forming an almost unbroken cover in spite of carrying a full complement of eggs. Incorporated with it was a little Senecio lautus, some dead, some green and flowering. Stunted Muehlenbeckia complexa occurred on vertical faces, terminating flush with the summit, where it became very foul with guano but escaped being trampled upon. On the most southerly of the three northern stacks the Disphyma-Senecio community was more open, the interstices occupied by extensive mats of dead Poa annua. This nitrophilous grass is a common winter annual in both British and New Zealand bird colonies, becoming established on the fertile soil when the birds leave and the guano concentration becomes suitable diluted by rain, and dying out with the return of the birds and higher evaporation rates of summer. The place of the Muehlenbeckia was taken by Coprosma repens, heavily fouled but thriving. The central northerly stack was divided into two parts, the higher supporting only partially dead Disphyma, the lower more open Disphyma with guano-covered but thriving Salicornia australis occupying about 5%. The most northerly colony resembled the second, except that the Senecio as well at the Poa was dead. Terns were observed nesting on fairly pure but worn, yellowish Disphyma with scattered Coprosma on Takapu or Passage Rocks west of Ponui Island. Others occurred with gulls on the Disphyma occupied ledges below the gannets on Motu Takupu Island off the west coast of the Coromandel Peninsula. Terns were found nesting with red billed gulls, small black shags, probably white-throated shags (Phalacrocorax melanoleucos) brevirostris (Gould, 1837) and little pied shags (Microcarbo melanoleucos) off the west end of Motuihe Island, mainly on three offshore stacks. Most of the outermost and lowest stack was used as a perching area and was devoid of vegetation, but a few terns' and gulls' nests occurred among sparse Disphyma in open crevices on the eastern end. The central stack carried an 80% cover of Disphyma with a few terns' nests Scattered among numerous gulls'nests.

The innermost stack was loftier and partially covered with 30–100 cm. high Coprosma repens in which the shags nested. Gulls and terns, with a preponderance of the former, nested on the remainder, principally on Disphyma, but a few straying on to a belt of the dead annual grass Avena fatua. Gulls and terns at the foot of the sandy cliffs of the main island were among Disphyma, stunted Coprosma and dead grass. A small tern colony of rather more than 150 pairs occurred on a stack off the north end of Motuhurakia Island in the Noises Group. The eggs were more widely spaced than on Crusoe Island, there being an average of 1½ m. between each pair, and some occurred under low Coprosma bushes. There was no dense Disphyma mat here, this species giving a coverage of only 10–70%. The area was liberally plastered with guano and, although the vegetation was less specialised and more diverse, only 8 species were recorded in the tern colony, 3 of them marginal, as against 20 on an adjacent part of the same slope of equal area. Spp. of Tern Colony Disphyma australe d. (Marginal spp.) Coprosma repens a. * Bromus sterilis.Anisantha sterilis. * Lepturus incurvatus or Pholiurus incurvus.Parapholis incurva (dead) f. Muehlenbeckia complexa Senecio lautus r. Salicornia australis Spergularia media r. Spp. of Rocks Unoccupied by Terns Disphyma australe a. Astelia banksiac r. Pittosporum crassifolium a. * Calamagrostis filiformis or Agrostis avenacea.Deyeuxia forsteri r. Parapholis incurva f. Mushlenbeckia complexa r Coprosma repens f. Orobanche sp. r. Carex lucida l.f. Salicornia australis r. Dichondra repens l.f Scirpus nodosus i. Asplenium flaccidum l.f. Sonchus oleraceus r. Auisantha sterilis o. Spergularia media r. Senecio lautus o. Vulpia sp. r. Anagallis arvensis r. Poa annua v.r. A larger colony occurred on David Rocks in the same group of islands, but the nests were again as much as 1½ m apart. The majority of the terns nested on the lower slopes on almost pure Disphyma or among 30 cm. high Coprosma. Others were higher up the clift where subordinate species included 3 grasses, Parapholis incurva and Vulpia sp. locally abundant and Avena fatua rare. Higher nests occurred under the edge of a mixed scrub of Metrosideros excelsa (= M. tomentosa), Pittosporum crassifolium and Nothopanax arboreum? This was fairly open bush with an undergrowth of Disphyma, Senecio, Vulpia and Avena. 3. White Island, Bay of Plenty Apart from a congregation of gulls and terns on a Disphyma covered stack off Rabbit Island, near Mount Maunganui, the only terns visited in the Bay of Plenty were those on White Island, 60 miles east of Tauranga. Terns nested in two places on the main islands, which was much larger than other tern island investigated, and also on tall stacks to the S. known as the Black Rocks. (Probably also on the 60 m high Volkners Stacks in the N. W.) White Island covers an area of approximately I square mile and rises to 1,053 et at Mt. Gisborne. It is an active volcano, reputed to function as a safety valve.

for the whole of the North Island thermal region.Much of its area is occupied by the crater, near which no vegetation can survive owing to frequent ground movements and dense sulphurous fumes. The corrosive nature of the atmosphere is demonstrated by the marked decomposition of the rock surfaces and the equipment left in the abandoned fertiliser works. Oliver (1915) suggested that the chief factor determining the peculiarities of the vegetation was the poisonous steam charged with HCl fumes which rose from the surface of the crater lake. This lake disappeared in the eruption of 1914, actually before his paper appeared in print, but there are still plenty of active steam vents. Fleming (1948) remarked that probably no part of the island was quite free from the effect of the permeation of acid gases.Such vegetation as survives is also subjected to a constant rain of fine volcanic dust which showers from leaves and branches at the slightest touch. The plants persisting under these condition are few and, like seabird vegetation, are predominantly halophytic although ascending 5–600 feet above sea level.In most cases their leaves are protected from salt spray, wind desiccation, guano, corrosive fumes and dust alike by a thick epidermis (e.g., Coprosma repens, Disphyma australe) a woolly tomentum (e.g., Metrosideros excelsa, which is exceptionally resistant in other areas to the guano of shags nesting in its branches) or scale-like hairs (e.g., Chenopodium allanii = C triandrum). These 4 species and the notably guano-tolerant Poa anceps v. condensata were those of chief importance, and only 4 others were recorded—viz., a few clumps of Phormium tenax on eastern clifftops, patches of Histiopteris incisa in dry shady creeks beneath trees in the west, a few clumps of the man-induced weed Phytolacca octandra in the vicinity of the fertiliser workers' old camp at Ohaurora and a single plant of Sonchus oleraceus in a gull-frequented part of one of the gannet colonies. Oliver (1915) recorded a total of 12 species, his list including Metrosideros robusta (habitat unstated), Deyeuxia forsteri and Gnaphalium luteo-album (“rare, shingle beaches”) and Solanum nigrum (“near gannets”). Of the 9 species recorded in 1957, only Phytolacca did not appear in his list. No vegetation was recorded on Captain Cook's chart, but in 1863 Captain Thomas reported that the island was covered on the N.and W.sides with a scrubby growth of timber and peopled by thousands of gannets and mutton birds. In 1924 “ there was at least 100 acres of pohutukawa forest capped by a cone of rose pink rock and broken up by white patches covered by thousands of gannets and seagulls” (White I. Agri.Chem. Co. Ltd. Bulletin, 1924). The “seagulls” may actually have been terns, as the only nesting gulls seen in 1957 were about 30 pairs of red-bills among driftwood in Crater Bay.Oliver (1915) makes no mention of either gulls or terns, only gannets and mutton birds. Pohutukawa (Metrosideros excelsa) was still the dominant vegetation in 1957, forming a double, horsehoe-shaped belt around the lower part of the cone; broken where the crater extended to the E coast at Crater Bay and Wilson's Bay. The lower part of the belt consisted of more mature forest, made up of trees 3–8 m.high. This was a fairly even-aged stand forming a closed community with practically no undergrowth and terminating abruptly at its upper margin. The upper part of the belt was again an even-aged stand but very open and of individuals only ½–¾ m.high protruding from loose rubble at 2–4 m. intervals. Contrary to appearance these young pohutukawas were not the pioneers of a primary sere but of a secondary sere, having become established on ground which had previously borne plants. The surface of the intervening pink and grey “boulders” could be chipped off to reveal that the rock-like appearance was only

Table I Percentage Frequency of 15 Species in 15 Tern Colonies in Bay of Plenty and Hauraki Gulf. November, 1957 Species Category Locality (The 5 annual grasses marked with an asterisk are aliens) Perennial Annual Halophyte Rabbit I. White I, Troup Head White I, E. Coast Passage Rocks Crusoe I. S. Crusoe I. N. S. Stack Crusoe I. N, Mid Stack Crusoe I. N, N. Stack Motuihe I. W. Stack Motuihe I, Mid Stack Motuihe I, E. Stack Motuihe I, W. Cliff Motuhurakia Stack David Rocks Gannet I Percentage Frequency Disphyma australe H P d d d d d d d d d d d d d d d 100% Coprosma repens H P o r o a a a a 47 Senecio lautus H P r o o r o 33 *Avena fatua A f f 13 *Parapholis incurva H A f f 13 *Poa annua A f o 13 Chenopodium allanii H P o o 13 Muehlenbeckia complexa H P r r 13 Salicornia australis H P r r 13 *Vulpia sp. A f 7 Metrosideros excelsa H P o 7 *Anisantha sterilis A o 7 Nothopanax arboreum? H P r 7 Pittosporum crassifolium H P r 7 Spergularia media H P r 7

skin-deep. Beneath a brittle crust ¼–1 cm. thick was a friable sulphur-yellow or iron-brown soil, richly impregnated with the fine fibrous roots of a former vegetation. It appeared that liquid mud spilling over the rim of the crater above may have solidified over the broken ground or caused a heat metamorphosis of the surface. That the lower, more mature belt of forest was a relict of a vegetation which previously extended higher was suggested by its survival on parts of the N. E. coast only on eminences which would have been by-passed by downward-flowing material from above. Some of the depressions between were still largely bare right to the cliff edge, supporting only small scattered bushes. None of the “pseudo-rocks” were found in the lower forest, which was floored by a considerable depth of peaty soil copiously tunnelled by thousands of greyfaced petrels (Pterodroma macroptera gouldii Hutton, 1869) and sooty shearwaters (Puffinus griseus). (Both mutton bird spp. recorded by Oliver (1930), only Pterodroma mentioned by this author in 1915.) The same type of “rock skin” was seen in the crater, but the soil beneath contained no plant remains. A zone of Disphyma australe of varying width skirted the lower fringe of the forest, a fair proportion of Chenopodium allanii incorporated with it, and at the junction of the two communities was another, generally a narrow belt of Coprosma repens, sometimes of Poa anceps. The main tern colonies were located on Troup Head between Crater Bay and Wilson's Bay and slightly further N. along the E. cliffs where the herb zone broadened. Both were shrouded in fumes when the volcano's smoke plume was blown down by a N. W. wind, but the birds congregated mainly on the seaward slopes 1–200 feet above sea level where the fumes were less than those endured by the red-billed gulls just above H.W.M in Crater Bay. (The only land bird seen on the island was a house sparrow (Passer domesticus Linnaeus, 1758) in the vicinity of the old camp—a species which has shown itself well adapted to withstand the sulphurous fumes of the great industrial cities of the world.) The tern rookeries were occupied by a 75% cover of Disphyma with frequent Chenopodium. When visited early in the breeding season (16–18 November) the plants were in healthy condition, most of the bare patches being rocky and uncolonisable. 4.Summary of Tern Vegetation and Comparison with that of Other Areas Table I shows the limited nature of the flora of the tern colonies visited, only 15 species being recorded in as many areas.Disphyma australe was the dominant plant in every case, Coprosma repens and Senecio lautus were present in less than half, and the other 12 species were found in only 1 or 2. Of these only 4 of the alien annual grasses, Avena, Parapholis, Poa and Vulpia rose to the category of “frequent”, three of them occurring more than once, in spite of the remoteness of the rookeries from human interference. Spray-washed cliffs offer an inhospitable plant habitat and the flora is always limited, but not to such a marked extent.Areas of similar aspect and exposure which were not subjected to the added rigours of guano deposition and trampling usually showed a more diverse flora than did the adjacent rookery. The shortage of woody species other than low-growing maritime shrubs was probably not significant, as it is doubtful if terns could kill these back. The birds, which have unusually fickle nesting habits, sometimes deserting a rookery one year and returning the next, would naturally select an area fairly free from scrub.Even.

on David Rocks there was little evidence that they were killing back the marginal bush and spreading on to the vacated ground as gannets and shags do. White fronted terns were seen to be quite abundant on the Stewart Island waters of south New Zealand in February, 1957, but no nesting colonies were visited.Large gatherings of birds congregated on rocks beneath the Half Moon Bay lighthouse and on Ulva Island and Goat Island in Paterson Inlet and the following plants, none of which were found in the north New Zealand tern colonies, were recorded:— Asplenium obtusatum * Crassula moschata.Tillaea moschata Poa astoni Hebe elliptica * Scirpus cernuus.Isolepis cernua Leontodon leysseri Scirpus nodosus Sonchus oleraceus Carex trifida Sonchus asper The absence of plants such as Disphyma, Salicornia, Senecio and Spergularia was largely due to the paucity of blown sea spray in the sheltered Stewart Island habitats; the absence of many of the other northern species was due to latitudinal differences between the flora of the “sub-antarctic” south and the “sub-tropical” north. A nesting colony of the slightly larger crested tern (Sterna bergii cristata (Stephens, 1826)) on a reef off Little Dog Island, in the Eastern Bass Strait, Tasmania, approximately 1,550 miles W.S.W.of the Hauraki Gulf colonies was, like them, dominated by Disphyma australe. Almost as common was Tetragonia implexicoma* “In his description of T. trigyna Banks & Sol ex Hook. f. TI Nov Zel Cheeseman (1906) writes.‘It is probably identical with the Australian T. implexicoma Hook f’ and if this is so as it is believed to be by Melbourne taxonomists. T. implexicoma has priority as the older name In the author's experience, there appear to be more differences between the Western Australian and S. E. Australian plants (both of which are included in T. implexicoma) than between the N. Z. and S. E. Australian ones which appear identical.” and Salicornia australis was quite frequent. These three species were the only members of the tern flora, although two others occurred on a part of the reef unaflected by the birds. The reef was visited towards the end of the nesting season (22.1.58) when the chicks were almost fully grown and the plants very worn, a considerable number having been killed. The quantity of freshly dead stems in even the foulest areas indicated either that the vegetation regenerated seasonally from portions which had survived the breeding season in the partial protection of crevices or that the reef had been recently colonised by the birds (Terns had not previously been recorded here by C.S.I.R.O.ornithologists doing intensive studies in the area during the previous 12 years.) “Disphyma australe was an important constituent of three other crested tern colonies in the Furneaux Group, Bass Strait (Scott's Reefs, Tuck's Reef and Nett's Reef.)” A much larger colony of crested terns nested on the N. coast of Cat Island on the Tasman Sea fringe of the E.Bass Strait Group in 1954–5, and its effects on the vegetation were still very apparent three years later, in February, 1958. There was no Disphyma, but the rocks previously occupied by terns were dominated by the related Carpobrotus rossii (= Mesembryanthemum aequilaterale) (28% ground cover), adjacent unoccupied rocks by Poa poaeformis (=Poa billardieri) (22%) and Pelargonium australe (16%) with Carpobrotus occupying only 1%. On the deeper soil of the more inland part of the colony Bromus unioloides (= B.catharticus) had invaded the old nest site to give a 30% cover.This species.

also dominated an area vacated by gannets on this island and the nesting grounds of Cape Barren geese on other Bass Strait islands; elsewhere on Cat Island it was rare or absent. In New Zealand it was commonly associated with gannets in the Hauraki Gulf and red billed gulls on the Mokohinau Group. The dominant Poa poaeformis on Cat Island dropped from 70% to 28% in the old tern colony and Sonchus oleraceus increased from negligible amounts to over 12%.Tetragonia implexicoma increased from 20–27%, this guano-loving species being generally abundant throughout in association with short-tailed shearwater burrows and not depending on the presence of surface nesting birds to such an extent as Bromus appeared to Tetragonia trigyna was not recorded in the tern or gannet colonies of northern New Zealand, but was found to be quite characteristic of penguin, shearwater, petrel and gull colonies from the Mokohinau Group in the N.to the Stewart Island Group in the S. In Western Australian tern colonies Disphyma was replaced as dominant by Carpobrotus rossii. Caspian terns (Hydroprogne caspia, Pallas, 1770) breed from New Zealand to Western Australia—e.gs., of N.New Zealand nesting sites including Bay of Plenty (Motiti Island), Hauraki Gulf (Crusoe Island) and Bay of Islands (Motu Arohia Island).In all three areas the caspian tern nests were only just above high water mark on the loose sand or shingle of a spit connecting two islets, where the substrate was too mobile and too close to sea level to support plant life. In Bass Strait they were most often associated with Disphyma, in Western Australia with Carpobrotus. II “Natural” Gannet Colonies 1. General Observations The Australian gannet (Sula bassana serrator (Gray, 1843)) is much more destructive of vegetation than is the tern, and in well-established colonies it is the rule rather than the exception for all macroscopic plants to be eliminated from the nesting area. Habitat requirements are as for the white-fronted tern, and the two species may share the same stack or the one take over from the other as on the Black Reef at Cape Kidnappers (Fleming and Wodzicki, 1952).These authors have shown that gannets are spreading in northern New Zealand, this entailing extension of the nesting areas, and in newly annexed territory selection of resistant plants is still in the early stages and a few coprophiles survive. These become excessively deep green around and particularly below the nests due to the high nitrogen content of the soil, and flowering is often inhibited, possibly by the low C: N ratio. The lush nitrogen-induced growth is more susceptible to mechanical damage than is normal growth, but certain species, notably Disphyma australe, survive in good condition in the initial season of bird colonisation until well after the chicks have hatched. These make good during the succeeding winters some of the damage suffered as long as the nests remain fairly widely spaced, but the tendency is for the gaps to become occupied by further pairs of birds and the plants are progressively ousted. Areas may, however, be vacated by the birds and thus made available for recolonisation by plants. Nests are often closer in bush colonies than in the open as outward expansion is restricted by the surrounding trees and may only proceed as fast as these are killed off by the guano at their roots. The leaves of the only two trees bordering gannet colonies at all frequently, Metrosideros excelsa and Coprosma repens, are very resistant to coatings of guano, this property being afforded by the external protective layers which conserve water and help to keep out salt spray. Oliver.

(1915) has described this xeromorphic leaf structure in relation to the resistance of the two species to volcanic fumes on White Island. Gannets taking off badly from this type of colony not infrequently make a crash landing in the bushes and flap themselves free only with difficulty, suggesting that the open habitat is likely to be preferred if available (as with the European gannet, Sula bassana (Linnaeus)). Bush is the natural climax vegetation, however, in northern New Zealand and an open habitat other than on coastal rocks or where gannets are already established, is usually maintained only by grazing, periodic burning or other human management.Generally gannets prefer to breed in areas more remote from the activities of man, but exist amicably enough side by side with farm livestock in the Cape Kidnappers Plateau colony. 2. Colonies in the Hauraki Gulf Horuhoru or Gannet Rock, off the N.E corner of Waiheke Island, in the.Hauraki Gulf, was visited on 29.11.57, when the chicks had attained almost full size. The island is 90ft high, topped with Coprosma repens bush and belted below by Disphyma australe. Gannets nested in both zones, and it was evident from the remains of dead trees scattered through the upper part of the colony that they had become established initially on the Disphyma zone and worked upwards as the trees died. The uppermost birds were very crowded, nesting close against the trees in places, and the branches of these were leafless and dead for more than than 1 m from the periphery of the bushed zone. The extent of inward migration of the gannets depended partly on the local topography and was irregular, the various centres of population being separated by belts of Coprosma and some consisting of as few as 20 nests. Considerable patches of the alien annual or biennial grass, Bromus unioloides, occurred fairly close to the birds.This was dead, as it was in most places at this time of year, so not necessarily as a result of gannet activity. Its survival as a “drought evader” was more vital here than in less specialised habitats, as it enabled the life cycle to be completed before a summer soil water deficit brought the guano in the soil to toxic concentration. The species is a characteristic coprophile of farmyards as well as being a common seasonal invader in bird colonies. Other conspicuous species on the western slopes where the gannets nested, but not close to them were a lush, green and inaccessible patch of what appeared to be Apium prostratum and a single bush of Lycium ferocissimum. Gannets did not occur on the steep eastern face of the island, but were present on the ridge above and the vegetation benefited from the diluted guano washed down by rain, and was unusually green. An overflow from the main colony nested on a small stack off the N.end of the island where the vegetation consisted solely of Disphyma. Breeding colonies on three of the islands of the Motu Kawao Group, off the W.coast of the Coromandel Peninsula were visited. These were on Gannet Island (Motu Takupu), Bush Island (Motu Karmarama) and a stack off the N.end ofDouble Island (Motu Wi). The first was almost bare of vegetation, with Disphyma in crevices below the gannets. The second was much larger and supported three main colonies on headlands jutting from the N.and N.W sides of the main bush-covered portion of the island.In addition there were small subsidiary colonies tucked back in clearings of the bush and two further ones on offshore stacks.

On the level areas the nests were close together and surrounded by Metrosideros excelsa bush from which they were separated by a belt of Coprosma repens. The lesser colonies in the clearings contained only 20–30 nests each, these occupying all the available clear area. All the larger colonies occurred on the lower, open slopes where the terrain was less even and the nests more widely spaced. As on Horuhoru, Disphyma and Bromus unioloides were the chief species, occupying much of the ground within the colony and forming the bases of the nests in places.In many areas the Bromus also formed a marginal belt between the gannets and the bush. Some Senecio lautus was present. The stack off Double Island supported a small colony of about 40 nests, among which were the woody remains of a former bush vegetation. Above was a strip of Coprosma separating the gannets from the Metrosideros of the stack summit, and below us was a lush flowering growth of Disphyma, patches of dead Bromus and a little Senecio. 3. White Island Colonies The White Island gannetries covered an area of approximately three acres (Grange, 1927) and occurred in three main areas— one on West Point in the N. W., three on Rocky Point, further E. along the S.coast, and two (or one divided by a gully) on Gannet Point, in the S.E. The first and last were on open promontories and expanding rapidly, the increase in the census figures for 1946–47 over 1926–27 being from c. 500 birds to 1,254 pairs on West Point and from c. 700 birds to 2,565 pairs on Gannet Point (Fleming & Wodzicki, 1952). The three Rocky Point colonies were smaller and surrounded by dense bush, the number of birds having been estimated at c. 700 birds in 1926–27, whilst in 1946–47 1,408 nests were counted but only 607 of them were occupied—this indicating a substantial increase followed by a decrease. In 1957 most existing nests were occupied, but there were areas which the birds had vacated where the guano-saturated ground was being gradually colonised by Disphyma australe, Chenopodium allanii and Poa anceps, the two latter only in the local shelter of rocks or partially dead branches. In the expanding colonies of the W.and S. E. Points the vegetation was withdrawing and many birds were occupying “virgin” habitats on thick mats of Disphyma, Chenopodium or Poa. The area was visited early in the nesting season (16–18 November, 1957), when some of the chicks had hatched but some pairs were still nest-building, and the plants had suffered relatively little damage as yet, although liberally covered with guano around the growing mounds of nest material. Grange (1927) states that the excrement of the birds was found only to a depth of 3ft, and that its fertilizing value proved to be lower than anticipated, with a nitrogen content of less than 0.05% and an average phosphoric anhydride content of 2 11% with a maximum surface reading of 4 42%. Its manurial content was sufficiently potent, however, to kill off the most coprophilous of plants where the nests were at all close. The relative distribution of gannets and vegetation is illustrated by sketch maps in Fig. 1. (a) On West Point it was evident that formerly isolated groups of nests were beginning to coalesce, many of the intervening sites having been occupied for the first time during the current year. The Disphyma adjacent to these was healthy but showing the effects of the high nitrogen content in the deep blue-green colour of the foliage.This contrasted sharply with the yellowish-green of normal plants

Figure 1 White Island, Bay of Plenty Sketch maps showing the relative distribution of nesting gannets and vegetation. 17. 11. 57.

when viewed from a distance, the difference being accentuated by the higher proportion of glaucous Chenopodium near the nests and the scarcity of flowers which might be attributable to the upsetting of the carbon: nitrogen ratio. There was plenty of room for further expansion free from tree cover, and it seemed that the position of some of the nests varied from year to year, some previously occupied sites remaining empty and allowing plants not completely killed to regenerate. This ensured a more uniform appearance in all but the oldest established parts of the colony, but the ground seemed not to remain free from gannets for long enough to permit the entry of shrubs. (b) On Gannet Point, the largest colony, there was also room for expansion, the Metrosideros excelsa forest being well back from the coast and much of the Coprosma repens belt being replaced by a luxuriant growth of Poa anceps. Some of the bare patches were steeply sloping, heavily burrowed by mutton birds and subject to erosion, but the presence of burrows alone was insufficient to deter the gannets and quite a few pairs were found sitting in the hollows formed by burrow entrances. In some cases the burrows had been effectively blocked by the accumulated nest material, and it is unlikely in any case that they could have remained occupied. This was most commonly seen on the Poa of the inland margin of the most westerly section of the gannetry where the birds were extending their range towards the trees at the expense of the mutton bird rookery. Naked and downy chicks and eggs were present, but many pairs were still building, plucking Disphyma, Poa and Chenopodium from around the nest and so hastening the degradation of the indigenous flora.(In the Rocky Point woodland colonies, where little or none of this material was available on the site, the nests were of seaweed, the species having been identified by Miss Moore (Moore & Wodzicki, 1950). Some birds were building on Disphyma-covered c.1 m. in diameter but many preferred natural hollows. The plants were surviving well, although photosynthesis must have been severely curtailed by the coating of guano. Disphyma and Poa were flowering, Chenopodium was in the vegetative condition, dark green and very large leaved. (c) On Rocky Point the gannet colonies were much nearer sea level than were the other two more open colonies and had eaten back into the Metrosideros bush in three places.Usually the coastal fringe of Coprosma was tolerant of more guano as well as more exposure than was Metrosideros, though it still protected the taller trees behind from wind-borne sea spray. Neither approached the gannets much closer than 2 m. and a fringe of dead wood separated the living parts of the trees from the birds. The death of the branches only on the side of the tree where the roots were subjected to guano suggests that there may be little lateral translocation of solutes from the excreta. These trees were more stunted than those suffering greater exposure but no guano deposition. The low cliffs below the gannets as well as the bush behind were tunnelled by mutton birds which might conceivably have occupied the entire area had they not been in competition with the larger birds. The Disphyma, Poa and Chenopodium occupying the seaward mutton bird rookery were creeping up into vacated parts of the gannetry. (d) General “Gannet Succession.” Disphyma was the pioneer, advancing closest to the gannets and trailing up over the marginal, wind-trimmed Coprosma for distances of 2 m. and to heights of 1½ m. It did not, however, penetrate far beneath their shade Chenopodium and Poa pioneered the areas beneath the dead branches and Poa penetrated furthest beneath the living bushes where light was sufficient. Coprosma seedlings were scattered in these herb mats.

It seems that the following sequence can be postulated for the degeneration of Metrosideros forest due to occupation by gannets and the regeneration of the forest as the gannets retreat:- Metrosideros. Coprosma. dead branches. bare ground. Disphyma. Chenopodium. and/or Poa. Coprosma. Metrosideros. The presence of mixed forest on the extinct volcano of Rangitoto Island, near Auckland, suggests that Metrosideros is not the true climax vegetation but that the succession is held at this stage on White Island by the prevalent volcanic fumes. 4. Black Reef Colony, Hawke's Bay In 1931, Black Reef was occupied by white fronted terns, the first record of gannets there being in 1938–39. For some years terns and gannets are said to have occupied the rocks in alternate breeding seasons (Fleming & Wodzicki, 1952). Mr. N. Elder, of Havelock North, states that a fair vegetation cover existed on the reefs during the early stages of the occupation. This included, among other species, Disphyma australe, Coprosma repens and Lepidium oleraceum, which latter is characteristically associated with island biid colonies from the Three Kings in the N. to Stewart Island in the S. The Lepidium and other species were killed out by the birds and the Disphyma and Coprosma were pushed back to form a marginal fringe, very little of which remained by 1957. By 1947 gannets were occupying ledges of the mainland cliff opposite the reef (Fleming & Wodzicki, 1952) and 10 years later the number of birds scattered over the cliffs here was considerable. They were fairly widely spaced and had affected the vegetation only locally. Most of them occurred among long grass in a community more akin to the Cape and Plateau colonies of Cape Kidnappers, except that the plants remained ungrazed by livestock. 5.Summary of “Natural” Gannet Vegetation and Comparison with that in Bass Strait Most of the gannet colonies visited were higher above sea level than were the tern colonies (notable exceptions being the Black Reef gannetry and White Island ternery). They were thus more remote from sea spray on the whole, but the expected increase in diversity of the flora was prevented by the heavier deposition of guano. It was the rule rather than the exception for them to be on bare ground instead of on the Disphyma mats so characteristic of the tern colonies. Eleven species were recorded in seven colonies (some of the latter multiple colonies and collectively covering more ground than the 9 multiple tern colonies listed in Table I under 15 headings). Seven of these species were recorded in both tern and gannet colonies. As with the terns, Disphyma australe was the dominant species with 100% frequency, Coprosma repens the next most abundant (71%) and Senecio lautus third in importance (43%) with, in the case of the gannets and the Bass Strait terns, Bromus unioloides The only annuals recorded were again aliens—opportunist invaders of the open, well manured habitats created by the birds (Table II) Only one gannet colony was visited in the Bass Strait, Tasmania this being on Cat Island, several hundred metres inland from the deserted crested tern colony. The gannet colony had previously been much larger and the vacated area was being colonised by Bromus unioloides and Tetragonia implexicoma with an advancing

Table II Percentage Frequency of Plants in Seven Gannet Colonies of Northern New Zealand Species Category Locality (Annual spp. marked with an asterisk are aliens) Halophyte Annual Percentage † Spp recorded on Three Kings Islands by Oliver (1948)Three Kings Black Reef White I Double I Bush I Gannet I Horuhoru Frequency Perennial Disphyma australe H P x x x x x x x 100 Coprosma repens H P x x x x x 71 *Bromus unioloides A x x x 43 Senecio lautus H A P x x x 45 Lepidium oleraceum H P x x 28 Chenopodium allanii H P x 14 Metrosideros excelsa H P x 14 Poa anceps v condensata H P x 14 Rhagodia triandra H P x 14 Salicornia australis H P x 14 *Sonchus oleraceus A x 14 margin of Spergulamia rubra and Bulbine semibarbata. Bromus occupied 39% of the ground, Tetragonia 29%. The surrounding territory was burrowed by thousands of short-tailed shearwaters and showed no Bromus and little Bulbine. Poa poaeformis tussock, which occupied only 15% of the old gannetry, here occupied 59%; Tetragonia only 18%. (Soil around the surviving gannets was bare.) Thus, in the Australian tussock grassland, as in the New Zealand bush, the presence of gannets had led to replacement of the indigenous vegetation by maritime succulents and the large alien Bromus. The presence of the succulent Bulbine with the Tetragonia instead of Disphyma and Salicornia as on the tern reef off Little Dog to the W, reflected the slighter exposure to sea spray. III. “Exotic” Gannetry of Cape Kidnappers 1. History of the Gannet Occupation Cape Kidnappers is the only mainland nesting site of the Australian gannet and the most southerly. Its non-insular position and proximity to grazing land has led to an entirely different type of vegetation from that found in the previously discussed colonies, and this seems to merit separate treatment. The development of the three sections of the colony has been discussed by Fleming and Wodzicki (1952) and Wodzicki and Robertson (1953) and the following facts are relevant to the understanding of floristic differences between the three areas.

The part of the colony on the cape proper is the oldest and is thought to have been established about 1850. In 1946–47 nearly 3,000 nests were counted there, but 600 of these were unoccupied. A hundred years of occupation destroyed the vegetation on most of the cape, and the only nearby survivors seen in 1957 were plants of Disphyma australe over the brink of the vertical northern cliffs. It is only at the junction of cape and plateau that birds and plants come into close proximity and this end of the gannet colony has been populated only since 1920, a large patch on the S. side of the W. slope even more recently (1932). The marginal vegetation may thus be regarded as having been influenced by gannets for approximately 30 years. Figure 2 Plant Zonation in Relation to Proximity of Gannets Cape Kidnappers plateau colony 8. 1. 57.

The plateau colony 500 yards S. W. of the cape is the most recent of the three. Nests first appeared between 1931 and 1944, but the area had been used for roosting for some years before that. Approximately 200 nests existed in 1950. The colony is on the clifftop at the E. edge of undulating grazing land. No specific information relating to pasture management can be obtained, but Col. Nielson, owner since 1925, states that no sowing of grasses near the gannets had been carried out by him nor, to his knowledge, at all. According to members of the field husbandry department at Massey Agricultural College, most of this part of the country was sown 50–80 years ago with Lolium perenne, Dactylis glomerata, Trifolium repens and T. partense. The Lolium and later the Dactylis succumbed to heavy grazing and Agrostis tenuis (some of which may have been sown) and Danthonia spp. assumed dominance in many areas. The Kidnappers plateau was top-dressed by hand twice in the 1930's and again (from the air) on the date of the author's first visit (7.1.57). This latter dressing was of lcwt per acre of superphosphate with a little clover seed. The third Kidnappers colony, Black Reef, has been little influenced by agriculture and is discussed with the “natural” communities. 2. Zonation OF Vegetation Adjacent TO Gannets Concentric zonation of plants round the guano-saturated soil of the nesting areas is best seen adjacent to the plateau colony. This colony has been expanding since 1944 at the expense of the surrounding vegetation, and Wodzicki and Robertson (1953) report the first tentative colonisation of the northern sub-colony by gannets in 1946. This new site was then separated from the old by 8–9 yards of bare soil and 2–3 yards of grass sward By. the 1949–50 season about 50 pairs were nesting on the new site, and this grass was killed, but in 1957 the main sward still approached closer to these nests than to the earlier established southern group. Distribution of the more abundant species surviving during the height of the 1956–57 breeding season is shown diagrammatically in Fig. 2. The species rising to a maximum in the heavily manured soil at the forward margin of vegetation were Hordeum murinum, Urtica urens and Amaranthus deflexus, all characteristic farmyard coprophiles. Species finding their optimum in the high fertility belt a few metres behind these guano-tolarant annuals were Lolium perenne, Medicago arabica, Geranium molle and Erodium moschatum. Although the Lolium is likely to have been one of the main constituents of the original sown pasture, it was only here in the fertile submarginal zone that it had been able to maintain dominance in competition with less nutritious grasses—in spite of the fact that farm livestock had free access to the area and grazed right to the edge of the gannetry. Beyond the influence of the birds the vegetation consisted of coarse, rather tussocky grassland through which low woody clumps of Muehlenbeckia complexa straggled. Important components of this sward were Agrostis tenuis, Anisantha (Bromus) sterilis, Bromus mollis, Vulpia bromoides and Trifolium glomeratum with patches of Scirpus nodosus and Mariscus ustulatus. Eight months later, at the beginning of the 1957–58 breeding season, when the plateau colony had been unoccupied by gannets for approximately four months, the vegetation was seen to have advanced centripetally towards the nests. Not only had the annual Hordeum and Urtica migrated closer, but a belt of Poa annua approximately 2 m. wide had appeared beyond them, reaching to within 1 m. of the outermost nests in the small northern sub-colony and 3 m. of those in the older southern sub-colony. Nine of the marginal nests of the northern group were as yet unoccupied by birds, and all of these contained a few young Poa plants. All.

the nests of the larger southern group had been re-occupied by birds, so it was not possible to see whether there had been any tentative colonisation by Poa during the winter there. In British bird colonies some of the Poa annua which invades in winter, survives through the summer, but in the drier, warmer Hawke's Bay climate it apparently dies seasonally, and it formed a large proportion of the peripheral belt of dead grass recorded in January, 1957 (Fig. 2). Lolium perenne is common in the “medium fertility zone” in Britain (where not grazed out by rabbits) and Agrostis tenuis, Bromus mollis, Vulpia bromoides, etc., on the poorer ground beyond (Gillham, 1955 a, b, 1956). Most of the other ornithocoprophilous Kidnapper's species, Urtica urens, Hordeum murinum, Geranium molle, etc., are also found in British seabird colonies. The winter plant zonation adjacent to the Kidnappers plateau colony is shown diagrammatically in Fig. 3. Figure 3 Relationship of Gannets and the Dominant Vegetation Cape Kidnappers plateau colony. 8. 9. 57. Another resemblance of the two similar habitats on opposite sides of the world is in the algal flora, Prasiola, much of it in the Hormidium stage, increasing in both British and New Zealand nesting grounds after the birds depart in autumn. P. crispa has been found more commonly in Britain, P. stipitata in New Zealand, and specimens collected by Mr. Beauglehole from the Lawrence Rock Gannetry off the south coast of Victoria proved to be the first record of P. stipitata for Australia. The author has since had Prasiola specimens determined from bird islands off North Tasmania and S.E. Victoria and on Macquarie I. in the sub-Antarctic. An analysis of the macroscopic summer vegetation bordering the Kidnappers plateau colony is compared with that 20 m. away in Table III. Table IV shows that species growing in the vicinity of both the cape and plateau gannets approach nearer to the latter by an average of just over 2 m. This is due.

Table III Floristic Analysis of Vegetation Bordering Cape Kidnappers Plateau Colony to Show Plant Zonation in Relation to Proximity of Gannets (Data from 50 half-metre valence squares at 12 m intervals along five strips parallel to the edge of the gannetry) January, 1957. Species Distance in Metres from Forward Margin of Closed Sward 0- 1.5- 3.0- 4.5- 20 0- 1.5- 3.0- 4.5- 20 0.5 2.0 3.5 5.0 0.5 2.0 3.5 5.0 Av. % Ground Cover % Frequency Bare ground 16.4 80 * Mostly Poa annua with Hordeum murinum and a little Lolium perenne Dead grass 21.9 25.6 24.8 26.2 90 100 100 100 Hordeum murinum 22.0 8.1 7.6 4.7 1.1 90 80 90 80 50 Lolium perenne 17.9 34.0 52.5 55.4 36.7 70 90 100 100 100 Urtica urens 8.6 1.7 0.2 40 30 10 Amaranthus deflexus 4.8 4.3 0.2 20 30 10 Medicago arabica 3.6 10.3 6.8 3.3 2.7 40 100 100 80 70 Geranium molle 1.9 3.2 4.3 1.8 0.6 30 50 100 60 50 Carduus tenuiflorus 1.5 0.6 0.4 1.2 0.3 10 20 10 40 20 Malea parviflora 1.2 10 Erodium moschatum 0.2 1.4 0.7 0.1 10 30 20 10 † Of the 22 species present only Muehlenbeckia complexa and possible Solanum nigrum are native to New Zealand (S. nigrum is a doubtful native according to Oliver)Solanum nigrum 0.4 20 Medicago lupulina 0.3 0.2 10 10 Bromus mollis 0.1 2.7 4.8 11.1 10 40 70 90 Trifolium glomeratum 2.0 8.2 20 70 Crisium culgare 0.4 0.2 10 10 Holcus lanatus 0.2 10 Agrostis tenuis 14.2 100 Anisantha sterilis 11.0 100 †Muehlenbeckia complexa 7.6 50 Vulpia bromoides 4.6 7.0 Trifolium dubium 0.8 5.0 Dactylis glomerata 0.4 1.0 Total No of Species 9 11 8 9 17

Table IV Table Showing Preponderance of Exotic Annuals in Soil Disturbed by Birds. Gannet Colony, Cape Kidnappers. Distance in feet of the nearest individuals of each species from the 30-year-old end of the 108-year-old Cape colony, and the 15-year-old Plateau colony. (Species occurring near both breeding areas are more distant from the older established one by an average of 10 feet). Ratio of native to introduced species = 1.11 Ratio of annuals to perennials = 4.2.1 Species Native Introduced Halophyte Annual or Biennial Perennial Distance from Gannets Cape Colony Plateau Colony ft cm ft cm Carduus tenuiflorus I A/B 4 122 1 30 Poa annua I A 5 152 Disphyma australe N II P 6 183 Hordeum murinum I A 6 183 1 30 Polycarpon tetraphyllum I A 6 183 Stellaria media I A 10 305 Lolium perenne I B (P) 10 305 1 30 Sonchus oleraceus I A 12 366 Urtica urens I A 14 427 1 30 Coronopus didymus I A 14 427 Geranium molle I A 14 427 4 122 Sonchus asper I A 16 488 Dactylis glomerata I P 16 488 3 91 Cirsium vulgare I B 16 488 4 122 Erodium moschatum I A/B 1 30 Medicago arabica I A 1 30 Amaranthus deflexus I A 1 30 Malva parviflora I A 3 91 Anisantha sterilis I A/B 6 183 Poa anceps N P 6 183 Trifolium glomeratum I A 6 183 Holcus lanatus I P 10 305 Bromus mollis I A/B 10 305 Medicago lupulina I A/B 10 305 Solanum nigrum ? ? A/B 10 305

not only to the relative ages of the two colonies but also to the local topography and proximity of the sown pasture. Plants approaching the inner end of the cape colony lie at a lower level than the gannets and receive the drainage waters from the gannetry. This applies also to plants on the east cliff below the plateau colony, but not to those on the western flank to which Figs. 2 and 3 apply. The ground there is fairly level and the grass sward suffers lateral seepage of excreta but little direct drainage. As seen earlier, the two major components of bird floras are halophytic perennials which survive when the birds are present and non-halophytic annuals which invade when they are absent. Most of the annual invaders, in both New Zealand and Bass Strait, are aliens, and these are common on the plateau grassland where they produce a readier source of seed for the invasion of the plateau colony than of the more remote cape colony. A possible further factor hindering their spread on the cape might be the greater degree of exposure there to wind-borne spray. 3. Floristic Composition of Vegetation Adjacent to Nesting Gannets Comparison of Tables IV, V and VI shows that the gannet flora is composed essentially of alien species and contains a high proportion of annuals. Of the 38 annual species listed for the area as a whole only three are indigenous to New Zealand, and one of these, Senecio lautus, quite often behaves as a perennial. Of the 30 perennials listed, 21 are indigenous and only 9 introduced. Twenty-five species were recorded growing within 5 m. of the gannets, and of these only two were native (excluding the doubtful Solanum nigrum), giving a native: alien ratio of 1: 11. One of the two natives, Disphyma australe was typical. Table V. Table Showing Proportion of Aliens to Natives and Annuals to Perennials in Species Approaching to Within 16–50 Feet of the Cape Gannet Colony. Ratio of native to introduced species = 1: 1.75 Ratio of annuals to perennials = 1. 75: 1 Species Native Introduced Halophyte Annual or Biennial Perennial Agrostis tenuis I P Capsella bursa-pastoris I A Cynosurus echinatus I A Epilobium nummularifolium N P Gnaphalium luteo-album N A Lagurus ovatus I A Lavatera cretica ? I A/B Malva nicaeensis I A Scirpus nodosus N P Spergularia media N H P Trifolium dubium I A

Table VI. Table Showing Proportion of Aliens to Native and Annuals to Perennials in Cliff Species Not Approaching Within 50 Feet of the Cape Gannet Colony. Ratio of native to introduced species = 1.2:1 Ratio of annuals to perennials = 1:2 Species Native Introduced Halophyte Annual or Biennial Perennial Adiantum diaphanum N P Agropyron repens I P Arundo conspicua N P Carex sp. N P Cassinia leptophylla N P Centaurea melitensis I A Centaurium pulchellum N A Cotula australis I P Cotula coronopifolia N H P Dichondra repens N P Erigeron canadensis I A Gnaphalium subrigidum N P Leptospermum scoparium N P Linum monogynum N P Lobelia anceps N H P Mariscus ustulatus N P Marrubium vulgare I P Medicago hispida I A Melilotus indica I A Phormium sp. N P Plantago coronopus I H A/B Polypogon monspeliensis A Polystichum richardii N P Samolus repens N I H P Senecio lautus N H A/B P Silene gallica I A Sporobolus capensis I P Trifolium repens I P Vicia sativa I A Vulpia bromoides I A Wahlenbergia gracilis agg N P

of all the more remote gannetries visited; Poa anceps, the other, of the White Island gannetries. Species recorded within 15 m. of the gannets but not closer than 5 m. gave a narrower native: alien ratio of 1:1.75. Species of the cliff and plateau vegetation not found within 15 m. of the gannets gave a native: alien ratio of 1. 2:1. The ratios of annuals to perennials in these three zones were 4.2:1, 1.75:1 and 1.2 respectively (Fig. 4..) Figure 4 Diagram Showing the Increase of Alien and Annual Plants in Vegetation Affected by Birds Cape Kidnappers gannet colony January 1957 These two sequences illustrate how short-lived alien species, the majority of them weeds of cultivation or pasture, are stimulated where birds disturb and enrich the soil, even when the birds are not themselves bringing seed as gulls do. Conditions suitable to plant growth may be restricted to as little as 4–5 winter months each year, sufficient only for the maturation of ephemeral plants, and the perennials which are killed out by the birds have little opportunity to become re-established. The boundary of the gannetry acts in effect as a nursery for the weed species, ensuring an adequate supply of seed from which migration back to the pasture could take place in the event of the deterioration of the sward due to over-grazing or drought. The mesophytic Kidnapper's perennials compete less adequately with the annual invaders than do the halo-coprophilous perennials of the island gannetries. It seems possible that the lack of guano-tolerant halopytes is due, not only to the less exposed situation of the mainland habitat, but also to the incidence of grazing. The only succulent halophytes recorded, Disphyma australe, Spergularia media and

* This species is not usually eaten if food is plentiful.Senecio lautus, are quite palatable to stock and found only on the more inaccessible cliff faces. Acknowledgments I am indebted to the University of New Zealand and the Australian Science and Industry Endowment Fund Trustees for research grants to help towards the expenses of this work and to the New Zealand Government Marine Department for the provision of boat transport to islands and stacks normally accessible by no other means. Also to members of the D.S.I.R., Botany Division and Massey College Field Husbandry Department and Mr. Carnahan for the identification of specimens; to Mr. Beauglehole for information relating to the Lawrence Rock gannetry; Messrs. Elder and Nielson for information relating to the Cape Kidnappers gannetry, and Messrs. Braithwaite and Burdon for assistance in visiting the Cape. References Fleming, C. A., 1948. Gypsum at White Island. N. Z. Jour. Sci. & Tech. 29, pp. 84–88. Fleming, C. A., and Wodzicki, K. A., 1952. A census of the gannet (Sula serrator) in New Zealand. Notornis. 5, 2. pp. 39–78. Gillham, M. E., 1955a. Ecology of the Pembs. Islands. III. Effects of grazing on the vegetation. J. Ecol. 43, 1. pp. 172–206. — 1955b. Some effects of the larger animals on the vegetation of Lundy I. Trans. Devon Ass. lxxxvii, pp. 205–229. — 1956. Ecology of the Pembs Islands. V. Manuring by the colonial seabirds and mammals. J. Ecol. 44, 2. pp. 429–454. Grange, L. I., 1927. White Island. D. S. I. R. 21st Ann. Rep. pp. 19–20. Moore, L. B. and Wodzicki, K. A., 1950. Plant material from gannets' nests. Notornis. 4, 1. pp. 12–13. Oliver, W. R. B., 1915. The vegetation of White Island. Jour. Linn. Soc. Bot. 43, pp. 41–47. — 1930. New Zealand Birds. Wellington. — 1948. The Flora of the Three Kings Islands. Rec. of Auck. Inst. & Mus. 3, 4. pp. 211–238. Wodzicki, K. A., and Robertson, F. H., 1953. Notes on the life history and population trends of the gannet (Sula serrator) at the plateau gannetry, Cape Kidnappers. Emu. 53, 2. pp. 152–168. White Island Agric. Chem. Co. Ltd. Bull. 1924. Mary E. Gillham, Ph.D. B.Sc. 209 Gunnersbury Park. Ealing, London, W. 5.