Phyllanthus, Sida, and the dubious Digitaria. Four, possibly as 

 many as eight, are indigenous. In 1965, one vine of the 

 indigenous Ipomoea pes-caprae was also found, but three 

 subsequent surveys failed to locate it. 



Species-Area Relationships 



The relationship between the numbers of plant species and 

 island size has long fascinated biologists (Fosberg, 1949; 

 Wiens, 1 962; MacArthur& Wilson, 1967; Whitehead & Jones, 

 1969), yet data from uninhabited islands is scant. The studies 

 from Kapingamarangi (Niering, 1956; Wiens, 1956) and 

 Aitutaki ( Stoddart & Gibbs, 1975 ) treat atolls with long histories 

 of human occupancy. SomeofthevillagesonKapingamarangi's 

 23 motus date to 1200 A.D. Aitutaki's 16 uninhabited motus 

 lie adjacent to a westernized volcanic island in an "almost- 

 atoll." People on both these atolls have profoundly influenced 

 their flora. 



Caroline provides an opportunity to compare the numbers 

 of species on motus of different sizes in an uninhabited atoll, 

 then to compare the results with Kapingamarangi, Aitutaki, 

 and uninhabited islands in the Line and Phoenix groups that 

 have no introduced species and have experienced minimal 

 human contact. 



Comparisons of Species-Area Relationships with Other 

 Atolls : Studies of Kapingamarangi (Niering, 1956 (contributed 

 greatly to theories of island biogeography (Mac Arthur & 

 Wilson, 1967). Because its motus coverthe same range of sizes 

 as Caroline, the two atolls might be expected to exhibit similar 

 patterns. However, their species-area relationships are 

 completely different. On Kapingamarangi, islets less than 

 1 .4 ha showed a constant, small number of species, after which 

 islets up to 100 ha showed a direct correlation of area with 

 numbers of species. On Caroline, a motu of 1 .4 ha supports 

 almost two-thirds of the total number of species, and plant 

 diversity on islets up to 107 ha shows only a slight, but not 

 necessarily steady, increase (Table 5). 



Species-area relationships on the motus of Aitutaki 

 (Stoddart & Gibbs, 1975, Figs. 33 and 34 of that paper) 

 conformed to the Caroline model; the number of species 

 increased only slightly on motus from 4 to 71 ha. Unfortunately, 

 Aitutaki had only one motu less than 1 .4 ha, so comparisons for 

 smaller islets cannot be made. The floras of all three atolls have 

 been impacted by man, but Caroline far less so than the others. 

 Much of the floral diversity on larger islets at Kapingamarangi 

 is derived from plants introduced by man and cannot be 

 considered normal. Caroline and Aitutaki provide much better 

 samples of natural plant species-area relations on atolls. 



Six islands in the Line and Phoenix groups (Maiden, 

 Starbuck, McKean, Phoenix, Vostok, Birnie) are uninhabited. 

 Their flora is entirely native. All are Caroline's "neighbors" in 

 an oceanic sense, and all except Vostok are dry , receiving about 

 750 mm (30") of rain yearly. They are old, essentially filled- 

 in atolls, containing hypersaline central lagoons or no lagoon 

 at all. Although the largest island (Maiden) has the greatest 

 diversity, there is only a very small linear increase in plant 

 species with increasing area (Table 8). Plant diversity is more 

 a function of climate (hot and dry) and distance from source 

 areas, than size, similar to the situation on Caroline. 



The Question of Fresh Water : The Kapingamarangi data 

 were analyzed with availability of fresh water in mind (Wiens, 

 1962;Whitehead&Jones, 1969). These authors suggested that 

 1 .4 ha is the threshold at which a freshwater lens can develop. 

 Below this size only halophytes can survive. They argue that, 

 as there are only a limited number of salt-tolerant species, the 

 floral composition on islets below 1 .4 ha is relatively constant. 

 On larger islets, species numbers increase in direct proportion 

 to land area, because permanent groundwater promotes the 

 survival of an increasing variety of nonhalophytic plants. 



The groundwater versus plant model does not apply to 

 depauperate Caroline for a number of reasons: first, the number 

 of plant species is not constant on islets below 1.4 ha: in fact, 

 species are added faster on motus from 0.02 to 1 .4 ha than any 

 other size range. 



Second, on Kapingamarangi, the number of species 

 increased in direct relation to islet size from 1.4 ha to 100 ha. 

 On Caroline, species numbers increased only slightly from 

 1.4 to 22 ha and exhibited another minor increase from 70 to 

 108 ha (see Fig. 31; Tables 5,6; and Ecological Succession 

 section). Thus, Caroline's data do not support the area- 

 diversity theory. 



Third, Whitehead & Jones (1969) argued that the flora on 

 "small" motus lacking a freshwater lens (i.e., < 1 .4 ha) consists 

 only of salt-tolerant strand species. This is not true on Caroline 

 (Table 6). In addition to harboring the usual strand species 

 (Tournefortia, Portulaca, Laportea, Heliotropium, Boerhavia, 

 Lepturus), Caroline' s "small" motus also support inland species 

 that are generally considered nonhalophytic (Pison ia, Morinda, 

 Achyranthes, Cordia, Phymatosorus). Either these latter five 

 species are moderately salt tolerant, or on Caroline the minimum 

 islet size with a freshwater lens is much less than 1 .4 ha, or both. 



Fourth, Whitehead & Jones (1969) postulated that the 

 nonhalophytic species are those that control overall species- 

 area associations. This may be a good generalization for less 

 remote islands but does not hold up for atolls with depauperate 

 floras (Table 6). For example, on Caroline the halophytic 

 Ipomoea macrantha, I. pes-caprae, Scaevola sericea, Sida 

 fallax, Lepidium bidentatum, Hibiscus tiliaceus, Thespesia 

 populnea. and Tribidus cistoides, which theoretically should 

 only occur as strand species on the smaller islets, occur only on 

 larger islets. In addition, several nonhalophytes (e.g., Morinda) 

 were found at Caroline on small motus where one might only 

 expect to find strand species. 



Fifth, the above authors do not mention bird-dispersal of 

 seeds, which is probably a factor that needs to be taken into 

 account on remote islands: at Caroline, Pisonia and Boerhavia 

 contribute to the floral diversity of islets from 0.2 ha to 108 ha. 



Sixth, Caroline does not have an assemblage of nonstrand 

 plants that only occur on larger motus; the only naturally 

 occurring, nonstrand plant is Psilotum. 



Seventh, the greatest factor complicating our understanding 

 of Kapingamarangi's natural evolutionary processes is the 

 presence of numerous exotics: of its 98 vascular plants, only 

 38 (39%) are indigenous. Its exotics include numerous weedy 

 herbs and food plants, which occupy gardens, abandoned 

 house sites, taro patches, and plantations (Cocos, Pandanus, 

 Artocarpus). These man-made habitats are particularly 



25 



