• Allee effect;
  • colonization;
  • dispersal;
  • endemism;
  • inbreeding;
  • island biogeography;
  • macroecology;
  • outbreeding;
  • Platypodidae;
  • Scolytinae;
  • species–area relationship


  1. Top of page
  2. Abstract
  3. Introduction
  7. Acknowledgments
  8. References
  • 1
     Inbreeding and parthenogenesis are especially frequent in colonizing species of plants and animals, and inbreeding in wood-boring species in the weevil families Scolytinae and Platypodidae is especially common on small islands. In order to study the relationship between colonization success, island attributes and mating system in these beetles, we analysed the relative proportions of inbreeders and outbreeders for 45 Pacific and Old World tropical islands plus two adjacent mainland sites, and scored islands for size, distance from nearest source population, and maximum altitude.
  • 2
     The numbers of wood-borer species decreased with decreasing island size, as expected; the degree of isolation and maximum island altitude had negligible effects on total species numbers.
  • 3
     Numbers of outbreeding species decreased more rapidly with island size than did those of inbreeders. Comparing species with similar ecology (e.g. ambrosia beetles) showed that this difference was best explained by differential success in colonization, rather than by differences in resource utilization or sampling biases. This conclusion was further supported by analyses of data from small islands, which suggested that outbreeding species have a higher degree of endemism and that inbreeding species are generally more widespread.
  • 4
     Recently established small populations necessarily go through a period of severe inbreeding, which should affect inbreeding species much less than outbreeding ones. In addition, non-genetic ecological and behavioural (‘Allee’) effects are also expected to reduce the success of outbreeding colonists much more than that of inbreeders: compared with inbreeders, outbreeders are expected to have slower growth rates, have greater difficulties with mate-location and be vulnerable to random extinction over a longer period.


  1. Top of page
  2. Abstract
  3. Introduction
  7. Acknowledgments
  8. References

Since Baker & Stebbins’s (1965) influential book on the genetics of colonizing species, numerous attempts have been made to generalize traits of good colonists (Sailer, 1978; Parsons, 1983; Simberloff, 1986; Cronk & Fuller, 1995; Williamson, 1996), in particular those on islands (Whittaker, 1998). Successful colonization and establishment on islands for any group of organisms will be influenced by their life histories and reproductive systems. For instance, many invasive plants exhibit self-compatibility, which assures fertilization after long-distance dispersal (e.g. Baker, 1955; Cronk & Fuller, 1995). Some island floras (e.g. Hawaii) have none the less higher than average proportions of dioecy, perhaps a result of successful long-term establishment due to the apparent lack of detrimental inbreeding in immigrant dioecious plants (Whittaker, 1998).

Studies of differential colonization abilities for animals with different reproductive systems have not yet received the same attention. However, introduced insects in the continental United States are clearly over-represented by the insect orders Hymenoptera, Thysanoptera and Homoptera (Simberloff, 1986). The two first groups, and the majority of all introduced Homoptera (scale insects and white-flies), are haplodiploid (Normark et al., 1999) where females are diploid and haploid males result from unfertilized eggs. Many haplodiploid species have strongly skewed offspring sex-ratios and practise extreme inbreeding (Wrensch & Ebbert, 1993), a definite advantage to colonizing organisms. It is generally believed that such organisms do not go through severe inbreeding depression (when close relatives are the only alternative for mating) during colonization because haplodiploidy, and especially regular inbreeding, reduces the potential for inbreeding depression by having effectively eliminated deleterious alleles early in the history of a lineage. That sib-mating species mate before dispersal also makes them independent of mate finding at new sites (Kirkendall, 1993). Population size is thus probably less relevant to the survival of inbreeding colonizing lineages as compared to outbreeding equivalents.

Wood-boring and bark-boring beetles (hereafter referred to as wood-boring for convenience) in the curculionid subfamily Scolytinae are one of only two groups of beetles in which haplodiploidy is known (the other being the bizarre monotypic family Micromalthidae, which also breeds in dead wood) (Mable & Otto, 1998). All haplodiploid scolytines are regular inbreeders (Kirkendall, 1993; Normark et al., 1999) and the haplodiploid genetic system enables females to control their offspring sex-ratios perfectly (Hamilton, 1967). Although other genetic systems are also known from the seven independent origins of regular inbreeding in Scolytinae (e.g. paternal genome elimination [Brun et al., 1995]), the largest clade of inbreeding species is strictly haplodiploid and includes the tribe Xyleborini and the inbreeding genera in the tribe Dryocoetini (Normark et al., 1999). When females of these species have not been fertilized by their brothers, they can produce haploid sons parthenogenetically, mate with the first-born and eat those remaining (Büchner, 1961). Hence, colonizing females of these taxa do not have to be inseminated before dispersal to ensure successful colonization.

Throughout tropical and temperate forests, scolytine beetles dominate the guild of wood-boring beetles that attack recently dead or dying trees. In the tropics this guild also includes beetles of the related family Platypodidae (Browne, 1961; Atkinson & Equihua-Martinez, 1986). Almost all species of Platypodidae and more than half of the tropical species of Scolytinae [including the entire species-rich tribe Xyleborini, and some smaller groups (Beaver, 1989a)] tunnel deep into wood and culture mutualistic yeast-like fungi for larval food (ambrosia beetles). Most of the remaining scolytine species breed in and feed upon phloem (true bark beetles), the pith of twigs, or in seeds. Other wood-boring beetle families contribute much less to the guild, or have very different ecologies, and will not be mentioned further in this paper.

Inbreeding scolytines are significantly more abundant (in terms of species) in tropical latitudes than outside the tropics, and for the Cryphalini and ambrosia beetles the proportions of inbreeding species are significantly higher on small tropical islands than on large islands and mainland regions at similar latitudes (Kirkendall, 1993: his Table 7.2 and Fig. 7.1). Here, we add more data and extend the analyses of tropical incest further, by examining the effects of area, isolation and of island elevation on species numbers. Are there any consistent differences between inbreeders and outbreeders for any of the geographical variables listed? If so, could this be due to different properties associated with different mating systems? To examine the possibility that inbreeding has evolved on the islands themselves, we investigated patterns of endemism for inbreeders and outbreeders. The group with the most widespread species should have the highest colonization potential of the two and perhaps indicate less intra-island speciation. We also examine whether widespread outbreeders on islands are disproportionately associated with plants used as crop cultivars, implying dispersal by human commerce.


Figure 7. Proportions of inbreeding (solid dots) and outbreeding (open dots) species that are endemic, occur on neighbouring islands (less than 1000 km apart), occur in a larger region (e.g. the Pacific) or are widespread (at least Africa + Asia). The eight islands were selected arbitrarily among those having four or more outbreeding species and with a recent review of the wood-boring beetle fauna. Proportions are reported for inbreeding and outbreeding species separately and are connected by lines for illustrative purposes only. See Table 1 for details on sample sizes.

Download figure to PowerPoint

Different islands will doubtless have different histories of evolution and of human disturbance (and transport). However, by analysing many island faunas together in a macroecological analysis, a unifying picture will probably emerge from the apparent noise resulting from limited sampling of individual islands, and help to explain the current island diversity of wood-boring beetles.


  1. Top of page
  2. Abstract
  3. Introduction
  7. Acknowledgments
  8. References

Study area

To avoid spurious results due to sampling error, we aimed to use as homogeneous a dataset as possible by assembling data from areas with similar climates. Most of the data points are from the Old World tropics, but we also included tropical rain forest islands from the Pacific rim of the Neotropics. Although islands off Africa were included, dry-climate islands such as the Canary Islands and Cape Verde Islands were excluded (see Kirkendall, 1993 for justification). The apparently dry Galapagos Islands were included, however, due to the extensive moist forest patches in part of the island group. Chichi-Jima (Bonin Island) has a tropical climate and was the only island included from outside the tropics of Cancer and Capricorn. The rather different faunal history of the Caribbean area, relative to the Pacific and the Old World where the largest group of inbreeders originated (Jordal et al., 2000), have led us to exclude islands from that region.

Taxonomy and sampling

Species lists were compiled from the literature and carefully checked for recent synonomies. Some unpublished records based on identifications made by RAB have been included for certain islands. Taxonomic imprecision is not as evident as for most other tropical insects because of the frequent attention paid to wood-boring beetles by forest entomologists and taxonomists. The designation of Scolytinae as inbreeders or outbreeders is based on recent reviews of inbreeding species (Kirkendall, 1993; Normark et al., 1999).

Bias in sampling effort among islands could produce spurious results. However, it is likely that uneven sampling will affect different groups equally, and samples from many localities are likely to reduce the remaining error. Bias resulting from differential ease of collection can be obviated by comparing inbreeding and outbreeding groups with similar ecology and hence with similar ‘apparency’ to collectors, in this case the xylem-boring and ambrosia-feeding Platypodidae and Xyleborini, or the inbreeding and outbreeding species of the predominantly bark and twig breeding Cryphalini.

Island sizes, altitudes and distances were obtained from The Times atlas of the world (The Times, 1988) and Douglas (1969). Since wood-boring habitats are necessarily within forests, forest cover is a much more precise estimate of area, however difficult to measure. Fortunately, forest cover correlates so closely with island size in SE Asia (Harcourt, 1999) that island area can be used as an index of available habitat. Nine islands larger than 30 000 km2 (such as Borneo, Madagascar and Sri Lanka) were treated as ‘mainland’ for analysing the effects of distance from potential colonizing sources. Single islands were included rather than archipelagos; the most species-rich island from each archipelago was selected as well as the second most species-rich island if it was more than 25 km away and if considerable collecting effort had been documented. The Philippines were treated as one ‘mainland’ island due to the group’s size and narrowly separated islands. In the Palau island group, the narrowly separated Babeldoap and Koror islands were treated as one island, as were the narrowly separated North, Middle and South Andaman groups.

Colonization abilities

Regressions were performed on log-log (x + 1) data, which usually provides the best fit to species–area data (Rosenzweig, 1995). We applied stepwise multiple regression to isolate the best predictor(s) from among area, isolation (distance) and altitude of the islands. To test whether regression slopes of the best predictor(s) were significantly different for inbreeding and outbreeding species, we used t-tests as described by Zar (1984). We assumed that the fauna of the mainland and larger islands represents the main pool of species available to colonize islands.

High colonization potential for a group should be reflected in a higher proportion of widespread species and a low proportion of endemics. For eight islands with a recent and precise beetle taxonomy, we classified each species as endemic (occurring on a single island), neighbouring (also on islands or mainland closer than 1000 km), regional (present within 1000–5000 km, e.g. Pacific, SE Asia or Africa), or widespread (present on at least two continents, Africa + Asia, or Asia + Oceania). The proportions of each category were calculated for inbreeding and outbreeding species separately.


  1. Top of page
  2. Abstract
  3. Introduction
  7. Acknowledgments
  8. References

Geographical predictors

Area was the single significant variable accounting for variation in the data (t = 4.56, P < 0.01, for islands 12–47, Table 1). Although island elevation was slightly correlated with island size (r = 0.37), it did not account for much variation in species numbers (t = 0.44, NS), nor did distance to the mainland (t = 0.39, NS) or nearest 10-fold-larger island (t = −0.96, NS). Changing the sequence of input variables to the stepwise regression did not alter the results, nor did inclusion of more homogeneous classes of data, singly or in combination: islands <1000 km2, <1000 m altitude or <1000 km distance to the ‘mainland’. Area was always the strongest predictor, but not always significantly so for such small subsamples. We therefore analyse only area effects in the remainder of this section.

Table 1.  (a and b) Species diversity of Scolytinae (and groups within) and Platypodidae on Pacific and Old World tropical islands and adjacent mainlands, with associated geographical variables. Areas 1–11 were treated as ‘mainland’ for the distance analysis. See METHODS for further details
Island and ‘mainland’Distance (km) to ‘mainland’Dist. to 10× larger islandMaximum altitude (m)Log area (km2)Total (Scolytinae + Platypodidae)ScolytinaePlatypodidaeInbreeding
  • (Austr.) Indian Ocean.

 1Borneo   *   *40005.879605421184299
 2Madagascar   *   *26005.769227187 40111
 3Sumatra   *   *38005.719272188 84140
 4Papua New Guinea   *   *50005.602681363318223
 5Malaya   *   *22005.522514380134290
 6Philippines   *   *30005.477354271 83180
 7Sulawesi   *   *35005.276247190 57128
 8Sri Lanka   *   *25004.817149132 17 83
 9New Britain   *   *23004.577124 89 35 61
10Taiwan   *   *40004.556149108 41 63
11Equatorial Guinea   *   *10004.449 80 52 28 34
12Viti Levu3300330013244.017102 95  7 61
13New Ireland 570  3818703.937 73 51 22 36
14Vanua Levu3300330010323.743 37 33  4 22
15Andaman 290 290 7383.660 51 44  7 36
16Reunion 700 70026003.398 13 11  2  9
17Bioko  30  3030003.301 43 30 13 27
18Mauritius 850 850 7703.301 29 27  2 22
19Oahu5300530010003.097 33 32  1 28
20Upolu4950110011003.049 60 56  4 36
21Tahiti6000240022373.018 22 21  1 15
22Santa Cruz 950 950 4002.954 19 19  0 15
23Grand Comore 300 30024002.942 10  8  2  7
24San Cristobal 900 900 5002.747 12 12  0 11
25Guam18601860 3942.733 16 15  1 11
26Babeldoap/Koror 900 900 2182.611 29 26  3 15
27Lanai5300 10010002.544 13 13  0 11
28Ponape16001200 7912.524 22 21  1 18
29Niue5000 880  702.413 14 13  1  9
30Tongatapu4000 705  822.410 10  9  1  5
31Hivaoa6000360012592.382  9  9  0  8
32Christmas Island* 350 350 9552.301 20 19  1 16
33Mahe13001300 9142.161 37 34  3 28
34Tutuila5000 100 6502.113 38 37  1 22
35Saipan1960 210 4742.086 17 17  0 11
36Tinian1920 280 1702.079 11 11  0  6
37Uapou6000360012312.017  9  9  0  6
38Kosrae16001200 6282.000 16 15  1 12
39Yap1200 450 1761.929 14 13  1  8
40Rarotonga69001400 6521.826 15 14  1    9
41Cocos Island 500 500 8501.668 19 19  0 14
42Praslin13001300 4271.544  5  4  1  4
43Tol16001170 4401.531  9  9  0  7
44Chichi-Jima10001000 3201.380 15 15  0 14
45Moen16001200 3701.279 11 11  0  7
46Aitutaki69001330 1241.255  8 8  0  5
47Niuatoputapu4200 275 1071.204  8 8  0  7
Island and ‘mainland’Out-breedingProportion inbreedingOutbreeding ambrosia beetlesXyleboriniXylo-sandrusScolyto-platypusDryocoetini inbreedingCryphalini outbreedingCryphalini inbreeding
 1Borneo3060.491922491584620 9
 2Madagascar1160.49 44 73 34131525
 3Sumatra1320.51 87105 632913 8
 4Papua New Guinea4580.33319175 614064 8
 5Malaya2240.56139224 75463119
 6Philippines1740.51 86144 63254511
 7Sulawesi1190.52 60 90 43262911
 8Sri Lanka 640.56 17 56 70182412
 9New Britain 630.49 35 47 201111 3
10Taiwan 860.42 46 50 75 9 7 4
11Equatorial Guinea 450.43 28 21 00 5 1 7
12Viti Levu 410.60  7 38 30112511
13New Ireland 370.49 22 34 10 0 6 1
14Vanua Levu 150.59  4 15 20 5 7 1
15Andaman 150.71  7 25 20 5 0 4
16Reunion  40.69  2  7 10 1 1 2
17Bioko 160.63 13 17 10 1 1 7
18Mauritius  70.76  2 15 40 4 4 3
19Oahu  50.85  1 18 20 3 4 6
20Upolu 240.60  4 24 40 613 6
21Tahiti  70.68  1  7 00 2 5 4
22Santa Cruz  40.79  0  6 10 2 0 7
23Grand Comore  30.70  2  6 00 0 0 1
24San Cristobal  10.92  0  4 00 2 0 6
25Guam  50.69  1  6 10 1 3 4
26Babeldoap/ Koror 140.52  3 10 10 2 7 3
27Lanai  20.85  0  9 10 0 2 2
28Ponape  40.82  1 12 10 2 1 4
29Niue  50.64  1  4 00 1 3 4
30Tongatapu  50.50  1  3 00 1 2 1
31Hivaoa  10.89  0  5 00 1 1 2
32Christmas Island*  40.80  1  8 30 2 3 6
33Mahe  90.76  3 12 4012 6 4
34Tutuila 160.58  1 13 20 410 5
35Saipan  60.65  0  3 10 1 5 7
36Tinian  50.55  0  1 00 0 5 5
37Uapou  30.67  0  2 00 0 4 3
38Kosrae  40.75  1  9 00 1 2 2
39Yap  60.57  1  4 00 2 4 2
40Rarotonga  50.60  1  6 10 1 3 3
41Cocos Island  50.74  0  9 00 1 0 3
42Praslin  10.80  1  2 00 1 0 1
43Tol  20.78  0  6 10 0 1 1
44Chichi-Jima  10.93  0  9 20 2 1 3
45Moen  40.64  0  6 00 1 2 0
46Aitutaki  30.63  0  2 00 1 3 1
47Niuatoputapu  10.88  0  1 00 2 0 4

Log-log species–area curves were linear and significant for all data combined (Fig. 1), for all inbreeding and outbreeding species separately (Fig. 2), as well as for inbreeding and outbreeding groups of ambrosia beetles (Fig. 3) and Cryphalini (Fig. 4). The slope for all species combined had a Z-value of 0.37, a relatively high value, although it is similar to that for SE Asian ants (Rosenzweig, 1995), for instance.


Figure 1. Species–area relationship for wood-boring beetles of Platypodidae and Scolytinae combined, in the Pacific and Old World tropics. See Table 1 for details on each data point. The linear relationship was highly significant (t = 16.5, P < 0.01) and justified further tests for equal slopes.

Download figure to PowerPoint


Figure 2. Test for equal slopes for all inbreeding species (solid dots, stippled line) and all outbreeding species (open dots, solid line) used in Fig. 1, respectively: slopes were significantly different (t = 7.62, P < 0.01, two-tailed).

Download figure to PowerPoint


Figure 3. Test for equal slopes for inbreeding (solid dots, stippled line) and outbreeding (open dots, solid line) ambrosia beetles: slopes were significantly different (t = 12.50, P < 0.01, two-tailed). Outbreeding ambrosia beetles include Platypodidae and Scolytoplatypus, while inbreeding species are all Xyleborini.

Download figure to PowerPoint


Figure 4. Test for equal slopes for inbreeding (solid dots, stippled line) and outbreeding (open dots, solid line) Cryphalini: slopes were significantly different (t = 5.67, P < 0.01, two-tailed).

Download figure to PowerPoint

Reproductive systems

Slopes of species–area plots were significantly different for inbreeding and outbreeding species, and slopes for outbreeding species were consistently steeper than those for inbreeding species (Figs 2, 3 and 4). The slopes diverged towards the left of the plots, implying that outbreeding species occur less often or that inbreeding species occur more frequently on small islands than expected from their occurrences in larger areas. This was especially clear from the comparison of subsets of ecologically similar groups with different mating systems (Figs 3 and 4). In particular, outbreeding ambrosia beetles (all Platypodidae +Scolytoplatypus) are absent or nearly so from the smallest islands. While the proportion of inbreeding species was clearly higher on small islands than on larger ones (Fig. 5), no association was found between the inbreeding ratio and distance from the source pool (Fig. 6).


Figure 5. Proportion of inbreeding species of Scolytinae and Platypodidae in relation to island area. The slope is highly significant (t = −5.41, P > 0.01).

Download figure to PowerPoint


Figure 6. Proportion of inbreeding species of Scolytinae and Platypodidae in relation to the distance from the mainland or islands larger than 30 000 km2. The result was similar when analyses were confined to islands smaller than 1000 km2, and lower than 1000 m maximum altitude, minimizing the effect of area and altitude.

Download figure to PowerPoint

Inbreeding species were more widespread than outbreeding species (Fig. 7). The latter had higher levels of endemism than inbreeding species in seven of the eight islands analysed. The exception, Oahu, has many inbreeding Hawaiian endemic Xyleborini, but there are very few outbreeding species on this most isolated island group.

Widespread outbreeding species

Although most outbreeding species found on islands have a more restricted distribution than inbreeding species (Fig. 7), there are a few exceptions. These species are either extreme generalists such as the platypodid ambrosia beetles Crossotarsus externedentatus Fairmaire or the recently pantropical Euplatypus parallelus (F.), or are associated with plants that are important crop cultivars. In the latter group, we find the pantropical mango beetle Hypocryphalus mangiferae (Stebbing), and the Pacific Ficicis porcatus (Chapuis) and Hypocryphalus mollis (Schedl), which are predominantly associated with Moraceae, in particular breadfruit (Beaver & Maddison, 1990). Almost all other outbreeding species, at least in the Pacific, are much more restricted in their distribution.


  1. Top of page
  2. Abstract
  3. Introduction
  7. Acknowledgments
  8. References

The area effect

Not only do inbreeders constitute a significantly higher proportion of species on small islands than on large islands and mainland areas (Kirkendall, 1993), but the proportion of inbreeders is inversely correlated with island size (Fig. 5). This contrasts strongly with the insignificant distance effect observed in this study (Fig. 6), an effect which often plays a significant role in the dynamics of island biota (Whittaker, 1998). Even if the island data are divided into more homogeneous classes (see legend, Fig. 6) we find no relationship between distance and proportion of inbreeders, or with total species diversity. Apparently, beyond some distance shorter than 1000 km, all islands are equally isolated with respect to bark and ambrosia beetle colonization (Fig. 6). Even islands relatively close to sources have higher proportions of inbreeding species than has the source area: Bioko Island (Fernando Pò: 63% inbreeders) is only 25 km from the coast of Cameroon and 160 km NW of Equatorial Guinea (43% inbreeders), Christmas Island (80%) is 400 km south of Sumatra (51%), and Grand Comore (70%) is only about 300 km east of Madagascar (49%). This suggests that even relatively short stretches of open water are enough to favour the establishment of inbreeding species. Neither did we find any additional effect of maximum island altitude, in contrast to the nearly equally important effect of altitude on the numbers of Hawaiian insect species per island (Peck et al., 1999). Our negative result for altitude might indicate that wood-boring insects undergo less intra-island speciation than most other insect groups, or that many island guilds are dominated by recent introductions.

A difference between outbreeders and inbreeders in the species–area relationship could be due to: (1) island sampling biases in favour of inbreeders; (2) major ecological differences between the taxa which inbreed and those which outbreed; or (3) that inbreeding facilitates colonization of smaller islands.


Sampling bias in favour of inbreeders only affects our analyses if it is different for small islands than for larger land areas. Such a bias could arise, for example, if inbreeding taxa are generally more abundant than outbreeding taxa, or if they are generally less host specific. In both cases, casual sampling (and collecting by non-specialists) could result in encountering inbreeding species disproportionately more often. However, this bias should apply to both larger and smaller areas alike, unless there are consistent differences in faunal composition due to ecological differences among inbreeders and outbreeders. Furthermore, repeated collections from small individual islands seem to yield higher rather than lower proportions of inbreeding species, as shown for the Galapagos Islands [64–79% (Schedl, 1974; Bright & Peck, 1998)] and Cocos Island [63–74% (Bright, 1982; Kirkendall & Jordal, unpublished)]. Intensive collecting efforts on the medium-sized Fiji revealed a stable proportion of around 60% inbreeders (Browne, 1974; Beaver, 1989b; Beaver, 1995), while the accumulation of collections from large islands shows a slightly decreasing proportion of inbreeding species, for instance Papua New Guinea (36% to 33%), Borneo (51% to 49%), and Sumatra (53% to 51%) (R.A. Beaver, unpublished records). Increased sampling, therefore, might reinforce the differences we found, and as such diminishes potential problems related to undersampling (Preston, 1948).

Resource availability

Our comparison of ecologically similar groups with inbreeding and outbreeding species falsifies the hypothesis that colonization differences are determined by differences in resource availability between small oceanic islands and larger continental islands. The strongest test is the comparison between inbreeding and outbreeding ambrosia beetles, because most are host-plant generalists and therefore less affected by host-plant selection (Browne, 1958). However, it is conceivable that available host-plant diameter could bias the colonization process if platypodids (all outbreeders) prefer larger diameter material than scolytines. For this to result in a bias towards inbreeding ambrosia beetles on smaller islands, large-diameter material would have to be limiting on such islands, which is unlikely. Since the treefall rate is estimated to be 5–8 trees (>10 cm d.b.h.) per ha per year for several tropical forests (Rankin-De-Merona et al., 1990), even the smallest islands used in this study would supply sufficient material to maintain healthy platypodid populations. That small islands are more susceptible to severe storms might also produce more treefalls than in more secluded forests. Also, many xyleborine species on small islands use mainly large-diameter logs (Browne, 1961), for instance the pantropical Xyleborus affinis Eichhoff, X. perforans (Wollaston) and X. ferrugineus (F.), suggesting ample supply of large-diameter material. Moreover, the common outbreeding ambrosia beetles in the genus Scolytoplatypus (Scolytinae) prefer small-diameter material (Browne, 1961; B.H. Jordal, personal observation), but are not recorded from any island smaller than 12 000 km2 (the narrowly separated Mindoro Island of the Philippines). In contrast, the inbreeding and thin-branch-breeding genus Xylosandrus is found on almost two-thirds of the islands in this study (see Table 1).

Cryphalini as a whole are over-represented on small islands (Kirkendall, 1993: his Fig. 7.2D), due perhaps to their preference for twigs and thin branches, which are more readily available in any forest locale. Their common occurrence in forest edges and other marginal habitats (Browne, 1961; Wood, 1982; Atkinson & Equihua, 1986) provides ready exposure to wind and may increase the rate of transport by humans between islands and the rate of establishment on new islands. This might explain the less dramatic effect of small islands on the number of cryphaline species that such islands can support. Despite their common occurrence, outbreeding cryphalines are still clearly under-represented on small islands compared to the inbreeding cryphalines (Fig. 4). Also, all other outbreeding groups of comparable ecology, for instance the tiny species of Acanthotomicus (≈95 spp.) and Cyrtogenius (≈105 spp.), are completely absent from small islands, which suggests that outbreeding twig breeders as a whole face many of the same problems during colonization as do other groups of outbreeders.

The inbreeding advantage

Having argued that neither sampling bias nor biological differences unrelated to inbreeding/outbreeding are likely to account for the differential success of inbreeders on small islands, we are left with the hypothesis that inbreeding itself has favourable consequences. However, there is one other option that first has to be considered. Theoretically, inbreeding could evolve repeatedly on islands as a by-product of bottleneck effects during early colonization by outbreeding immigrants and thus increase the relative number of inbreeding species on islands. If so, there should be a multitude of outbreeding sister-groups remaining in the source area. There are no known examples of such mainland (outbreeding) — island (inbreeding) relationships in the Scolytinae. Rather, 99% of the inbreeding species belong to two large clades (1400 + 200 species, respectively) suggesting the relative rarity of inbreeding origins (Kirkendall, 1993; Normark et al., 1999). Taken together with the much wider distribution of inbreeding species relative to outbreeders (Fig. 7), only differences in colonization abilities between inbreeders and outbreeders remains as a plausible explanation for the preponderance of inbreeding species on tropical islands.

The favourable consequences of inbreeding must necessarily be related to the establishment of propagules, and not to their dispersal. We assume that wind dispersal over oceans is very similar in most flying insects of comparable size classes (1–10 mm), as well as for species transported in plant material. Survival after wind transport seems relatively high for at least some groups of insects, as illustrated convincingly by the high number of non-resident moths (of Australian origin) light-trapped on Norfolk Island (Holloway, 1996). Thus, there should be no difference between the proportions of inbreeding and outbreeding arrivals.

The likelihood of success of colonizing species will depend strongly on the number of initial propagules, the length of time the colonizing population remains small, and the frequency of colonization attempts. Outbreeders will be much more affected by all three than will inbreeders, to the extent that they are more severely affected by density dependent ecological and behavioural factors (e.g. the ‘Allee’ effect: Allee, 1931; Courchamp et al., 1999). The most serious problem facing recently established small populations may be that of demographic stochasticity: that is, random fluctuations in family size and survivorship which can readily lead to extinction of very small populations (Gilpin & Soulé, 1986; Lande, 1988). Contributing to the problems faced by outbreeders (but not inbreeders) is that of mate location, which also will be most difficult at low densities. The length of time a colonizing lineage will be at risk to stochastic extinctions is determined partly by the intrinsic rate of increase for the species. Because inbreeders have higher rates than do outbreeders, due to their highly female-biased sex ratios and similar family sizes compared to outbreeders (e.g. Browne, 1961; Kirkendall, 1993), inbreeders should be vulnerable to stochastic extinctions for a shorter period of time.

Another well-documented problem for small populations of outbreeding plants and animals, although perhaps of lesser importance (Lande, 1988), is the problem of inbreeding depression. During most colonizing events, populations will be small and will inbreed, and few lineages are likely to survive (Lynch & Walsh, 1998). One reason that larger islands have a relatively higher proportion of outbreeding species (especially ambrosia beetles, see Fig. 3), therefore, may be that multiple colonizations are more likely for larger than smaller islands. Larger islands present larger ‘targets’ and hence probably receive dispersing propagules more frequently and in larger numbers, both of which would increase the chances that outbreeding species would survive the effects of demographic stochasticity and inbreeding depression.

Widespread inbreeders

Given that evolutionary divergence is expected and frequently observed for island plants and animals, the extremely widespread distribution of many inbred species is striking and deserves further attention. Although a thorough discussion is beyond the scope of this paper, we will mention briefly the most obvious hypotheses: (1) inbreeders have very limited genetic variability and founding populations even less, and therefore new colonies are largely unaffected by genetic drift or selection and hence remain similar to source populations; (2) inbreeding populations are more likely to be recent colonizations, and hence have had little time to differentiate (e.g. Andreev et al., 1998); (3) inbreeders have relatively higher colonization rates providing more gene flow among islands, leading to a greater similarity of forms among islands than seen in outbreeding taxa; and (4) recent taxonomists have been less apt to describe (equally divergent) island inbreeders as new species, because it is notoriously difficult to establish species boundaries in inbreeding complexes in the Scolytinae (Wood, 1954, 1982). These four hypotheses are not mutually exclusive. However, we note that (2), alone, is unlikely to be widely applicable — there is no evidence that inbreeders as a whole have spread more recently than outbreeders, and we point out that many small islands are populated almost solely by extreme inbreeders. It is difficult to evaluate the importance of (4) until molecular data for widespread inbreeding taxa become available.

Molecular data are crucial to understanding patterns of dispersal and establishment, and the time-scales during which these patterns took form. While we have presented herein a somewhat coarse-grained biogeographic picture of wood-boring beetles that vary in their mating systems, further insight into the genetic make-up of such insects will contribute considerably to our understanding of their differential colonization abilities.


  1. Top of page
  2. Abstract
  3. Introduction
  7. Acknowledgments
  8. References

BHJ was supported by grant no. 123588/410 from the Norwegian Research Council.


  1. Top of page
  2. Abstract
  3. Introduction
  7. Acknowledgments
  8. References
  • Allee, W.C. (1931) Animal aggregations. University of Chicago Press, Chicago.
  • Andreev, D., Breilid, H., Kirkendall, L., Brun, L.O., Ffrench-Constant, R.H. (1998) Lack of nucleotide variability in a beetle pest with extreme inbreeding. Insect Molecular Biology, 7, 197200.
  • Atkinson, T.H. & Equihua, A. (1986) Biology of the Scolytidae and Platypodidae (Coleoptera) in a tropical deciduous forest at Chamela, Jalisco, Mexico. Florida Entomologist, 69, 303310.
  • Atkinson, T.H. & Equihua-Martinez, A. (1986) Biology of bark and ambrosia beetles (Coleoptera: Scolytidae and Platypodidae) of a tropical rain forest in Southeastern Mexico with an annotated checklist of species. Annals of the Entomological Society of America, 79, 414423.
  • Baker, H.G. (1955) Self-compatibility and establishment after ‘long-distance’ dispersal. Evolution, 9, 347349.
  • Baker, H.G. & Stebbins, G.L. (1965) The genetics of colonizing species. Academic Press, New York.
  • Beaver, R.A. (1989a) Insect–fungus relationships in the bark and ambrosia beetles. Insect–Fungus Interactions (ed. by N.Wilding, N.M.Collins, P.M.Hammond and J.F.Webber), pp. 121143. Academic Press, London.
  • Beaver, R.A. (1989b) Bark and ambrosia beetles (Coleoptera: Scolytidae) newly recorded from Fiji, and their potential economic importance. South Pacific Journal of the Natural Sciences, 9, 17.
  • Beaver, R.A. (1995) Additions and corrections to the bark and ambrosia beetle fauna of Fiji (Coleoptera: Scolytidae). South Pacific Journal of Natural Science, 14, 1126.
  • Beaver, R.A. & Maddison, P.A. (1990) The bark and ambrosia beetles of the Cook islands and Niue (Coleoptera: Scolytidae and Platypodidae). Journal of Natural History, 24, 13651375.
  • Bright, D.E. (1982) Scolytidae (Coleoptera) from the Cocos Islands, Costa Rica, with description of one new species. Coleopterist’s Bulletin, 36, 127130.
  • Bright, D.E. & Peck, S.B. (1998) Scolytidae from the Galapagos Islands, Equador, with descriptions of four new species, new distribution records, and a key to species. Koleopterologische Rundschau, 68, 233252.
  • Browne, F.G. (1958) Some aspects of host selection among ambrosia beetles in the humid tropics of south-east Asia. Malayan Forest Records, 21, 164182.
  • Browne, F.G. (1961) The biology of Malayan Scolytidae and Platypodidae. Malayan Forest Records, 22, 1255.
  • Browne, F.G. (1974) A summary of the scolytid fauna (Coleoptera) of Fiji, with some new species. Commonwealth Forestry Review, 53, 6371.
  • Brun, L.O., Stuart, J.J., Gaudichon, V., Aronstein, K., French-Constant, R.H. (1995) Functional haploidy: a novel mechanism for the spread of insecticide resistance in an important international insect pest. Proceedings of the National Academy of Sciences, USA, 92, 98619865.
  • Büchner, P. (1961) Endosymbiosestudien an Ipiden. I. Die gattung Coccotrypes. Zeitschrift für Morphologie und Okologie Tiere, 50, 180.
  • Courchamp, F., Clutton-Brock, T., Grenfell, B. (1999) Inverse density dependence and the Allee effect. Trends in Ecology and Evolution, 14, 405410.
  • Cronk, Q.C.B. & Fuller, J.L. (1995) Plant invaders: the threat to natural ecosystems. p. 241. Chapman & Hall, London.
  • Douglas, G. (1969) Checklist of Pacific Oceanic Islands. Micronesica, 5, 327463.
  • Gilpin, M.P. & Soulé, M.E. (1986) Minimum viable populations: processes of species extinction. Conservation biology: the science of scarcity and diversity (ed. by M.E.Soulé), pp. 1935. Sinauer Associates Inc., Sunderland, MA.
  • Hamilton, W.D. (1967) Extraordinary sex ratios. Science, 156, 477488.
  • Harcourt, A.H. (1999) Biogeographic relationships of primates on South-East Asian islands. Global Ecology and Biogeography, 8, 5561.
  • Holloway, J.D. (1996) The Lepidoptera of Norfolk Island, actual and potential, their origins and dynamics. The origin and evolution of Pacific Island biotas, New Guinea to Eastern Polynesia: patterns and processes (ed. by A.Keast and S.E.Miller), pp. 123151. SPB Academic Publishing, Amsterdam.
  • Jordal, B.H., Normark, B.B., Farrell, B.D. (2000) Evolutionary radiation of an inbreeding haplodiploid beetle lineage (Curculionidae, Scolytidae). Biological Journal of the Linnean Society, 71, in press.
  • Kirkendall, L.R. (1993) Ecology and evolution of biased sex ratios in bark and ambrosia beetles. Evolution and diversity of sex ratio: insects and mites (ed. by D.L.Wrensch and M.A.Ebbert), pp. 235345. Chapman & Hall, New York.
  • Lande, R. (1988) Genetics and demography in biological conservation. Science, 241, 14551460.
  • Lynch, M. & Walsh, B. (1998) Genetics and analysis of quantitative traits. Sinauer Associates, Sunderland, MA.
  • Mable, B.K. & Otto, S.P. (1998) The evolution of life cycles with haploid and diploid phases. Bioessays, 20, 453462.
  • Normark, B.B., Jordal, B.H., Farrell, B.D. (1999) Origin of a haplodiploid beetle lineage. Proceedings of the Royal Society of London, Series B, 266, 22532259.
  • Parsons, P.A. (1983) The evolutionary biology of colonizing species. Cambridge University Press, Cambridge.
  • Peck, S.B., Wigfull, P., Nishida, G. (1999) Physical correlates of insular species diversity: the insects of the Hawaiian islands. Annals of the Entomological Society of America, 92, 529536.
  • Preston, F.W. (1948) The commonness, and rarity, of species. Ecology, 48, 254283.
  • Rankin-De-Merona, J.M., Hutchings, H., Lovejoy, T.E. (1990) Tree mortality and recruitment over a five-year period in undisturbed upland rainforest of the Central Amazon. Four neotropical rainforests (ed. by A.H.Gentry), pp. 573584. Yale University Press, New Haven.
  • Rosenzweig, M.L. (1995) Species diversity in space and time. p. 436. Cambridge University Press, Cambridge.
  • Sailer, R.I. (1978) Our immigrant insect fauna. Bulletin of the Entomological Society of America, 24, 311.
  • Schedl, K.E. (1974) Scolytidae from the Galapagos and Cocos Islands. 289th contribution to the morphology and taxonomy of the Scolytoidea. Studies on Neotropical Fauna and Environment, 9, 4753.
  • Simberloff, D. (1986) Introduced insects: a biogeographic and systematic perspective. Ecology of ecological invasions of North America and Hawaii (ed. by H.A.Mooney & J.A.Drake), pp. 326. Springer-Verlag, New York.
  • The Times (1988) Atlas of the world. Times Books Limited, London.
  • Whittaker, R.J. (1998) Island biogeography: ecology, evolution, and conservation. Oxford University Press, Oxford.
  • Williamson, M. (1996) Biological invasions. Chapman & Hall, London.
  • Wood, S.L. (1954) A revision of North American Cryphalini (Scolytidae, Coleoptera). University of Kansas Science Bulletin, 36, 9591089.
  • Wood, S.L. (1982) The bark and ambrosia beetles of North and Central America (Coleoptera: Scolytidae), a taxonomic monograph. Great Basin Naturalist Memoirs, 6, 11359.
  • Wrensch, D.L. & Ebbert, M.A. (1993) Evolution and diversity of sex ratio: insects and mites. Chapman & Hall, New York.
  • Zar, J.H. (1984) Biostatistical analysis. 2nd edn. Prentice Hall, Inc., New Jersey.