NOTES ON THE SEABIRDS OF ROUND ISLAND, MAURIT'IUS FRANK B. GILL, CHRISTIANJOUANIN, AND ROBERTW. STORER As oneof the few seabirdbreeding stations in the westerncentral Indian Ocean,Round Island, 14 milesnorth of Mauritius Island, providesnesting sitesfor two tropic-birds(Phaethon lepturus and P. rubricauda), a shear- water (Puffinus pacificus), and the only known Indian Ocean breeding colony of the South Trinidad Petrel (Pterodroma arminjoniana). The paucity of informationon the biologyof the petrel, which was only dis- coveredon Round Island in 1949 (Murphy and Pennoyer,1952) despite numerousprevious visits by naturalists,prompts us to record here the observationswe made during three visits to Round Island in the fall of 1964. We also includea few noteson the other three, more widespread, Indian Ocean species. Landing difficulties usually limit accessto Round Island to October and November when the seas are calmest, but even then trips can be planned only on a day-to-day basis. Furthermore,field work on the island is productiveonly during the few morninghours before the intense middayheat developsand one'senergy is expendedin climbingthe steep, unshaded hillsides. Our visits to Round Island were as follows: 16 October 1964 from 07:00 to 10:00 (Gill), 22 October 1964 from 09:30 to 12:30 (Gill and Storer), and 21 November1964 from 09:00 to 12:00 (Jouanin). DESCRIPTION The 372 acresof volcanicrock formingRound Island rise to a maximum elevation of 860 feet and were once covered with a lush forest that included severalendemic palms (Mascarenarevaughnii and Dictyspermaalbum), a screwpine(Pandanus vandermeerschi), a fan palm (Latania loddigesii), as well as typical Mauritian evergreenforest hardwood elements. Well- known to herpetologists,these forestsonce supportedtwo endemic (to Round Island itself) geckosas well as a relict subfamily(Bolyerinae) of bold snakes.In the last 100 yearsravages by the introducedgoats and rabbitsthat now aboundhave reducedthe island'saspect to one of barren rock with only scatteredpalms and clumpsof vegetation(Figure 1). Erosionand cyclonedamage also contribute to the losso.f natural vegeta- tion. Round Island's curiousherpetofauna has suffered along with the vegetationand in addition, when calm seas permit, the avifauna is regularlyplundered by local fishermenwho prize the seabirdshighly as a sourceof freshmeat. Fortunatelyof the four seabirdspecies inhabiting the island, the South Trinidad Petrel is said to be the least appetizing. Re- cently both Vinson (196.5) and Newman and Bannister (1965) have 514 The Auk, 87: 514-521. July 1970 July 1970] Round Island Seabirds 515 Figure 1. Remnants of the original vegetation on the south slope of Round Island. emphasizedthe marked deterioration of the island's flora and fauna and the need for protection. THE SOUTH TRINIDAD PETREL The South Trinidad Petrel (Pterodroma arminjoniana) occurs at sub- tropical latitudes in the Atlantic, Indian, and Pacific Oceans. Two subspe- cies,P. a. arminjonianain the SouthAtlantic and the smallerP. a. keraldica in the Pacific are recognizable,the Indian Ocean populationbeing similar in size to P. a. arminjoniana(Murphy and Pennoyer,1952). These petrels are so conspicuouson Round Island that we find their retarded discoveryrather surprising. Throughout the morning they fly actively back and forth over the island and certainly attract one'sattention with their eerie calls. Vocal activity was greatestduring the early morning hours (07:00-08:00) and decreasedgradually thereafter, though a few birds could be heard up to midday. The calls are normally associated with aerial activities involving pairs and, as far as we could tell, were always given by the individual that was following closely behind the other during a "chase." Occasionallya bird on its nest would call, particularly if disturbed. The call we heard most frequentlywas a rapid seriesof up to twenty "ki" syllablesincreasing in intensity followedby an extendedseries .516 Gx•x., JOT..TAxN'I•,Ai•TD STORER [Auk, Vol. 87 .4 ß 1 1.6 T I M E (seconds) Figure 2. Spectrogram(using wide band-passfilter) of a South Trinidad Petrel's call recorded on Round Island. First 14 elements are a series of rapid "ki's" that increase in intensity and show some structural transition to the terminal "k-lu's" (last four elements). A complete call normally includes many more "k-lu's" than are shown here. of lower, melodicoscillating "k-lu" notes (Figure 2). On severaloccasions we heard a rather different, harsh, descending"kree kree kree kree kree kree." During the two Octobervisits, Gill and Storerestimated that a maximum of 75 pairs of petrels were using the island. Banding studiesare now neededto determine the extent of turnover of individual petrels actually on the islandat any given time and thus to ascertainthe total population size.We foundmost of thepetrels near the summitof the island(Figure 3), thoughsmall groups of between5 and 10 pairswere also nesting in several shelteredspots on the upper slopes. As are other populationsof this species,the SouthTrinidad Petrelson Round Island are highly variable in plumagecolor. In order to estimate the relative proportionsof the different phasesin the field, we divided the complexcontinuum of colorationinto three major catagories: 1) en- tirely dark brown, 2) dark throat and upper breast with white belly, 3) white below (occasionallywith gray vermiculations)including throat. Entirely dark birds were the most numerousof these forms and white- throated forms the rarest. The populationwas composedapproximately of 6:3:1 ratio of all dark forms to dark-throated forms to white-throated forms. Possiblefuture changesin this compositionwill be of interest. Correlatedwith plumagephase are differencesin foot color, light-phase birds having bicoloredfeet with the outer half black and basal half and tarsuspink. The feet of dark-phasebirds were entirely black while those of intermediateplumage types contained variable amountsof black on the tarsi and inner portions of the webbing. July 1970] Round Island Seabirds 517 Fibre 3. Pe•el n•ting a•a on the summit of Round Isl•d; Serpent Island is in the •ckground. Also correspondingto plumagecolor is variation in the underwingcolora- tion, a character for which Murphy and Pennoyer (1952) described species-spedficpatterns that might be important for speciesrecognition and the maintenanceof reproductiveisolation. Like the rest of the plumage the underwingcolor ranges from completelydark to predominantelywhite with a dark anterior border. It would be interestingto know whether such variability is found only in thosepopulations that are not sympatricwith a congener. Pterodroma arminjoniana is known to be a surface-nestingpetrel, and we found no evidenceto the contrary on Round Island. However, the nestswere almost always sheltered, sometimes under succulent plants or by tufts of grass,but usually under overhangingeroded rocks (Figure 4) or amonglarge surfaceboulders. They are invariably on the shadedsides of such structuresand thus protectedfrom the intensemidday sun. Except for a few pebblesand somesmall piecesof plants or bones,the nests consistonly of bare rock or dusty soil. On all three visits we found nestswith eggsor young of various ages, including nearly full-size downy young. The downy young (Figure 5) are light gray, almost whitish in color, and not the dark brown or dusky gray Murphy and Pennoyer(1952) describeas characteristicof this and 518 Gr•, Jout•m, •o $TORER [Auk, Vol. 87 Figure 4. Sheltered nes•g •te of a •rk-phase South T•idad Petrel on Ro•d Island. other speciesof Pterodroma. On 16 Octoberwe countedat least six downy young, somevery small, but noted no eggs. On 22 October we found 3 eggsand over 10 young,some nearly full-sizeand somevery small. On 21 NovemberJouanin examined six nests,three of which containeda single chick about 40 days old. The other three nestscontained one egg each, one of which was fresh. In addition many birds were always sitting in nest-like situationsbut without eggsor young. Newton (1958) found 10 eggsin October 1954 and 1 egg and 14 chicksin December1954. Of particular interest are the recent observationsby Jean-Michel Vinson (in litt.) of both incubatedeggs and full-sized young petrels on Round Island between 22 July and 9 August 1968. Thus it seemsthat some breedingoccurs in this populationof South Trinidad Petrels throughout the year, for if the timesof incubation(i.e. at least 50 days) and fledging (90+ days) are at all similar to those of Pterodromahadtata cakow (Wingate in Palmer, 1962), eggsproducing full-sized young by 22 July were probably laid at the end of March and the young from eggsfound fresh on 22 Novemberwould not fledge until late February. Populations of Pterodromaarminjoniana in the Pacific region and South Atlantic are known to breed throughoutthe year (Murphy, 1936; Murphy and Pen- noyer, 1952). July 1970] Round Island Seabirds 519 Figure 5. Adult South Trinidad Petrel (light phase) and downy young. At present we have no evidenceof migratory movementsby these petrels; elsewherethe speciesis known to be sedentaryand to occupy breedingstations throughout the year (Murphy, 1936; Murphy and Pen- noyer, 1952). The only at-sea observationsof South Trinidad Petrels in the Indian Ocean are by Newton (1958) in the waters around Mauritius and northward towards the CargadosCarajos archipelago. OTHER SEABIRDS Wherever soil existedin any quantity on Round Island, it was riddied with burrowscontaining Wedge-tailed Shearwaters. We also found shear- waters in rock