An unlikely survivor: the peculiar natural history of the Raso Lark Paul F. Donald and M. de L. Brooke

ABSTRACT The Raso Lark Alauda razae is one of the rarest and most threatened birds in the world. Until recently, almost nothing was known of its population, behaviour or ecology. Recent research has shown that the species exhibits a number of fascinating behavioural and ecological peculiarities.The tiny population, currently fewer than 150 individuals, is dominated by males, which outnumber females by two to one.The males differ from the females in bill structure and feeding behaviour, possibly representing an adaptation to reduce competition between the sexes in a hostile environment.The species’ population is correlated with rainfall, and numbers in the past must have fallen to extremely low levels, possibly even into single figures, during periodic long droughts. Nesting success is very low, owing to heavy predation of eggs by the near-endemic Cape Verde Giant Gecko Tarentola gigas, itself a threatened species. Despite this, and a growing number of other threats, the species still manages to maintain a precarious toehold on its tiny, drought-prone islet home, thanks largely to the continued absence of the introduced mammalian predators that infest neighbouring islands. Recent tourist developments on neighbouring islands, leading to greater visitor numbers, and a predicted increase in drought through climate change pose new and increasing threats. Urgent measures are required to secure the long-term survival of this remarkable and unlikely survivor.

n 28th April 1897, the young English southern half, however, he soon realised that the naturalist and explorer Boyd Alexander island was inhabited not only by abundant O(1873–1910) scrambled up the low seabirds, but also by an unfamiliar lark, which cliffs that fringe the southern rim of the tiny, he described as ‘so tame that we could have waterless and uninhabited island of Raso, one knocked many over with sticks’ (Alexander of the smallest of the Cape Verde Islands. Off- 1898 a,b). He compensated for the absence of shore, his ship, ‘a fine old American pilot-boat sticks on this treeless island by shooting a good of nearly eighty tons’, was forced by the absence number of them instead, and 29 of these speci- of a safe anchorage to beat backwards and for- mens still reside in the skin collections of the wards, waiting for him and his companions. British Museum of Natural History in Tring Alexander had previously visited several of the and the American Museum of Natural History other islands in this (at the time) Portuguese in New York. Alexander named his discovery colony, and can hardly have expected to make Spizocorys razae, without explaining why he his greatest discovery on one of the smallest and assigned it to a genus represented at that time least hospitable of the group. After emerging by a single southern African species. After onto the flat, low plain that makes up Raso’s spending taxonomic time in the monotypic

420 © British Birds 99 • August 2006 • 420–430 The peculiar natural history of the Raso Lark

(and now defunct) Razocorys, and then in collection. After another long absence from the Calandrella, the Raso Lark was finally recog- attention of ornithologists, the Raso Lark nised by Hall (1963) to be closest to the skylarks briefly reappeared in the writings of Mein- Alauda, a position recently confirmed by DNA ertzhagen (1951) and Bourne (1955), before the analysis (Keith Barnes pers. comm.) and sup- Abbé de Naurois embarked upon a series of ported by detailed behavioural observations visits to the island in the 1960s. David and Mary (Donald et al. 2003). The main structural differ- Bannerman, authors of the first book on Cape ences between Raso Lark and Sky Lark Alauda Verde’s birds (Bannerman & Bannerman 1968), arvensis (smaller size, less pointed wing and preferred to view Raso from the comfort of a longer, heavier bill of the former) may all result passing ship and did not land, instead extolling from adaptations to life on a small, arid island the bravery and hardiness of de Naurois and (Hall 1963; Burton 1971; Hazevoet 1995), and others in landing and staying on the island. the two species are closely related. Further specimens were collected until at least A year after Alexander’s discovery, the Italian 1970 (eight specimens of this date reside in naturalist Leonardo Fea (1852–1903) visited Lisbon Museum), but little was recorded of the Raso and collected a further 30 specimens, now species’ numbers, ecology or behaviour until in the Museo Civico di Storia Naturale, Genova. the late 1980s, and the standard account given (Fea was also, incidentally, perhaps the last nat- in BWP is based largely on early and sketchy uralist to observe alive the remarkable Cape accounts. Verde Giant Skink Macroscincus coctei,once endemic to Raso and nearby Branco and prob- Distribution, population and sex ratio ably hunted to extinction around a century Cape Verde is the southernmost of a number of ago.) The lark was then left in peace until 1922, northeastern Atlantic island groups known col- when José Correia (1881–1954), an Azorean lectively as Macaronesia, which also contains the naturalist, collected specimens for the American rather better known (at least ornithologically) Museum of Natural History in New York. Two Canary Islands, Azores and Madeira. It is the years later, 16 further specimens were collected only politically independent member of this during an expedition in the sailing ship Blossom group, having ceded from Portugal in 1975. The by the Cleveland Museum of Natural History, archipelago lies in the northeast trade-wind belt, Ohio, some of which are also in the New York at about 15°N and approximately 500 km west of Paul Donald Paul 207. Raso in October 2001. Recent rainfall resulted in a growth of green cover and most birds were nesting. The sandy areas in the foreground contained large numbers of holes, excavated mainly by male Raso Larks Alauda razae to extract small bulbs from below the surface.

British Birds 99 • August 2006 • 420–430 421 The peculiar natural history of the Raso Lark

Senegal, and consists of nine inhabited islands equally isolated Warsangli Linnet Carduelis and a number of smaller islets, the island group johannis may owe its existence on the opposite having a total land area of 4,033 km2, scattered side of the African continent to a similar over approximately 60,000 km2 of ocean. Raso process of postglacial stranding in a habitat lies in the northern group of islands, the so- with few competitors. It is perhaps no wonder called ‘Windward Islands’ (Ilheus do Barlavento), that, despite noting several similarities to the and is one of three uninhabited islands (the Sky Lark (including its call), Boyd Alexander other two are Branco and Santa Luzía) that lie did not consider the possibility of his new bird between São Nicolau and São Vicente. being closely related to that familiar but geo- Where did the Raso Lark, so closely related graphically distant species. to the Sky Lark yet separated from the nearest Given these alternative scenarios, the time of population of that species by more than 2,000 arrival of the Raso Lark’s ancestors in Cape km of sea and desert, come from? There are at Verde is unclear. In conformity with the idea of least two possible explanations. The first is that late Pleistocene speciation, Hazevoet (1995) and this population represents the last remnant of a others have suggested that the species’ arrival is species that was formerly far more extensive likely to have been recent, within the last across Africa, perhaps a western counterpart of 100,000 years. Ongoing genetic studies, the Oriental Lark A. gulgula, which replaces the however, indicate a difference between Sky Lark Sky Lark in southern Asia today. This may have and Raso Lark in the mitochondrial become extinct on the mainland as climatic cytochrome-b gene of around 7.2%. A conven- conditions changed, but persisted in Cape tional estimate is that avian cytochrome-b DNA Verde. Alternatively, the species may have diverges by approximately 2% per million years. evolved in situ on Cape Verde following coloni- Although the application of this standard has sation by Sky Larks (or their ancestors) during been severely criticised, a 7.2% divergence one of the Ice Ages, when the Palearctic fauna between Sky Lark and Raso Lark strongly sug- was pushed southwards into what is now the gests a divergence at least 2–3 million years ago, Sahara. As the ice retreated and the Sahara rather than during the glacial and interglacial reverted to desert, the Sky Larks gradually cycles of the late Pleistocene in the last few returned northwards, leaving some stragglers hundred thousand years (Keith Barnes pers. stranded behind on Cape Verde to evolve into comm.). This new evidence suggests that either the species that survives there today. The the Raso Lark and its ancestors must have sur- vived on Cape Verde for far longer than originally thought, or the Raso Lark began its separation from the Sky Lark elsewhere before colonising the Cape Verde Islands. There is no written or physical evidence of the species from any of the other Cape Verde islands, but unless it is an extremely recent arrival (within the last 20,000 years, which now seems highly unlikely), its ancestors would have been present when Raso was intermittently connected through low sea levels to the neighbouring islands of Branco (3 km2), Santa Luzía (35 km2) and São Vicente Michael Brooke 2 208. Raso Lark Alauda razae, Raso, Cape Verde, December 2002. (227 km ). It is therefore

422 British Birds 99 • August 2006 • 420–430 The peculiar natural history of the Raso Lark likely to have had a far larger distribution than it few birds are found here during the breeding has today, as the resulting island (which would season, but up to half the population moves of course have included all the submerged areas here when the birds are not breeding. At such now separating the present islands) may have times, all the birds present in this part of the extended to several thousand square kilometres. island form a single tight flock, and are more Populations may have persisted on several wary and less approachable than those else- islands after rising sea levels separated them, where on the island, possibly because of perhaps until the islands’ discovery and occupa- the presence of ‘Neglected Kestrels’ Falco (tin- tion by Portuguese settlers (and their accompa- nunculus) neglectus in the area. The remaining nying cats, dogs, rats and goats) from 1462 birds are scattered in pairs or small groups onwards. There is certainly plenty of apparently around the lava plain, with a large concentra- suitable habitat on the larger (but cat-infested) tion in the southwest of the island. Less than Santa Luzía, and claims that sub-fossil bones half of Raso’s 7-km2 area is regularly used by the found on São Vicente may be of Raso Lark are larks, and the species has one of the smallest intriguing (Donald et al. 2005). It may be no coinci- dence that the Raso Lark survives on the largest island in Cape Verde that has never been permanently inhabited by people. Raso is divided fairly equally into two landscape types. The first comprises a low, flat plain of decom- posing lava and tufa, inter- sected by small, dry ribeiras © Natural History Museum,Tring (stream beds, where water flows only infrequently) with a scant cover of grasses Paul Donald Paul and herbs. It is here that the 209. Raso Lark Alauda razae (left) and Sky Lark A. arvensis (right).Although majority of Raso Larks the Raso Lark is only around two-thirds the size of a Sky Lark, the bill is subsist, although birds are considerably larger, representing an adaptation to life on an arid island. not spread evenly across it. The highest densities occur along the tops of the low cliffs in the southwest of the island and along the ribeiras. Within these restricted areas, densities of birds are actually rather high for a lark, members of the Alau- didae tending to occur in low densities relative to other songbirds (Donald

2004), and the species © Natural History Museum,Tring appears generally more gre- garious throughout the year than other Alauda larks. The Paul Donald Paul other half of the island com- 210. Male (left) and female (right) Raso Lark Alauda razae. Males have prises hilly outcrops rising significantly larger bills than females, and obtain significantly more of their to 164 m, valleys and raised food by digging than the smaller females, which instead feed mostly by plains, and much is almost picking seeds and invertebrates from the surface and from under stones. The difference in feeding methods between the sexes may contribute to devoid of vegetation. Very the species’ unbalanced sex ratio.

British Birds 99 • August 2006 • 420–430 423 The peculiar natural history of the Raso Lark

ranges of any bird in the world. lations fluctuate greatly and their size is dictated There is little historical information on pop- largely by rainfall, which is a prerequisite for ulation size. Bourne (1955), writing of a visit in breeding. Catastrophic droughts lasting up to 1951, said that the birds ‘swarm’ upon the island 18 years have occurred throughout Cape Verde’s and are ‘totally fearless’. When the Abbé de recorded history (as recently as the 1940s, tens Naurois’ first visited Raso in 1962, he is said to of thousands of Cape Verdeans died in a have found ‘an abundance of adults and imma- drought-induced famine). After rain, the Raso tures awaiting him’ (Bannerman & Bannerman Lark population increases; after droughts it can 1968). On subsequent visits, however, the same fall to extremely low levels. This pattern is naturalist noticed decreasing numbers, apparent even from the few existing population recording only approximately 50 pairs in 1965. estimates, and there is a strong positive correla- By the early 1980s, the population appears to tion between population size and rainfall have fallen to as few as 10 pairs, though these (Donald et al. 2003). In the early 1980s, fol- estimates were not based on systematic surveys. lowing a drought of more than a decade, popu- The first systematic count of the species did not lations were extremely low. Whether any occur until 1986, when Hazevoet (1989) esti- long-term trend underlies these greatly fluctu- mated the population at 75–100 pairs. The pop- ating population estimates is unclear, but there ulation had clearly declined by 1996, when are reasons to suspect that there has been a Bakker & van Dijk (1996), in a two-day visit, long-term decline in the population. Rainfall found the species ‘surprisingly hard to find’ but patterns in the Sahel (fig. 1), which are likely to eventually managed to see around 30 birds. be a reasonable indicator of rainfall on Cape Other population estimates are presented in Verde (T. Spencer pers. comm.), show that at Ratcliffe et al. (1999) and Donald et al. (2003, the time of Bourne’s ‘swarm’ in the early 1950s, 2005). In October 2001, the population stood at and of de Naurois’ ‘abundance’ in the early around 130 individuals. By November 2003, this 1960s, rainfall in the Sahel was well above had fallen to between 76 and 87 individuals, average. Since 1970, however, annual rainfall in and in December 2004 only around 65 birds the Sahel has generally been below the long- were present. However, following good rainfall term average, and populations may be lower on in 2005, the population increased sharply and average as a result. in December 2005, it stood at around 130 indi- Whether or not there has been any long- viduals. From these counts, it is clear that popu- term population decline, the Raso Lark is Michael Brooke 005. Raso Lark Alauda razae nest (centre, under overhanging vegetation), Raso, Cape Verde, December 2004.

424 British Birds 99 • August 2006 • 420–430 The peculiar natural history of the Raso Lark

3.0

2.5

2.0

1.5

1.0

0.5

0

-0.5

-1.0

-1.5

-2.0

-2.5 1898 1904 1910 1916 1922 1928 1934 1940 1946 1952 1958 1964 1970 1976 1982 1988 1994 2000 Fig. 1. Standardised Sahel rainfall index (average of June to October, the main rainfall season), 1898 to 2004. Dark blue bars indicate years with rainfall above the long-term average, light blue bars years with below average rainfall. Between 1970 and 2004, only seven years had rainfall that was at or above the long-term average. This drying trend, at least partly the result of increased greenhouse gasses, is expected to worsen in the future (Held et al. 2005) and poses a severe threat to the Raso Lark Alauda razae. Rainfall data reproduced courtesy of the Joint Institute for the Study of Atmosphere and Ocean (JISAO), University of Washington. steering a course perilously close to extinction. 2001 suggested, unexpectedly, that the sex ratio Species with greatly fluctuating populations are was actually far from balanced, and that males more at risk of extinction than species with greatly outnumbered females (Donald et al. stable populations, simply because there is a 2003). This was confirmed from biometrics of higher risk of chance events reducing the popu- birds caught for ringing in 2002 and 2003 lation to zero. When the population is as small (Donald et al. 2005); males outnumbered as that of the Raso Lark, extinction becomes a females by around two to one. This means that real possibility. That the species has managed to the total reproductive population is consider- survive for so long suggests remarkable ably lower than the actual population. The endurance, but it must have flirted with extinc- reasons for this skewed sex ratio are unknown, tion many times during its history, as the but may be related to differences between the extremely low counts from the 1980s suggest. sexes in morphology and feeding, discussed If the sex ratio among the 80 or so birds later. Such an unbalanced sex ratio clearly present in 2003 were one male to one female, increases the likelihood of extinction, since the population would equate to around 40 natural fluctuations need only to reduce the pairs. Most birds are assumed to have a sex ratio tiny number of females to zero for the species to that approximates to equality: females and become functionally extinct. The males’ songs males occur in equal numbers. However, might echo over Raso for several years after the finding an unbiased way to determine this in species becomes consigned to history through the field is extremely difficult. Males and the death of the last female. females may occupy different areas or habitats, may show a different propensity to enter traps Breeding ecology and display or nets and may be more or less conspicuous at The song, described in detail (with sonograms) different times of year, all making precise esti- by Hazevoet (1989), bears some structural simi- mation of sex ratios in wild bird populations larities to that of the Sky Lark. During pro- extremely difficult. In the case of the Raso Lark, tracted song flights, the song of the Raso Lark is however, estimation of the sex ratio is relatively more broken, slower and less complex than that straightforward: most of the world population of the Sky Lark, consisting of short bursts of can be seen within a single day, the species is four to six notes, little more than a repetition of tame, and males and females are usually easy to the call, interspersed with gaps. In all these distinguish in the field. Counts undertaken in respects, it resembles more closely the song of

British Birds 99 • August 2006 • 420–430 425 The peculiar natural history of the Raso Lark

the Short-toed Lark Calandrella brachydactyla. following irregular falls of rain, usually from However, the Raso Lark reveals its true ancestry August to April. Mass breeding events are prob- during the rapid descent at the end of the song ably less than annual, but it may be that some flight, when the song changes briefly to pairs are capable of breeding during any year, as resemble very closely the frenetic character of suggested by the presence of a recently fledged Sky Lark song. Song flights, given only by the juvenile in January 2003, following a long males, last an average of just over two minutes, period with no appreciable rainfall. Indeed, it is rather shorter than those of the Sky Lark, hard to imagine how the species could survive a though unpaired males give significantly longer drought of over a decade without some repro- song flights (up to a recorded maximum of 13 ductive input to the population. In October minutes) than paired birds (Donald et al. 2003). 2001 and again in November 2003, expeditions Singing birds face into the prevailing northeast to Raso to study the species fortuitously fol- trade wind and, in comparison with Sky Larks, lowed rainfall, and we now have data from over birds do not usually ascend as high during their 60 nests. song flights. On windless days and on days with The nest has been described and pho- strong winds, aerial song ceases and is replaced tographed by previous authors (Hazevoet 1989; by ground song, which appears more frequent Castell 1999) and closely resembles that of the than in the Sky Lark. As with Sky Larks, aerial Sky Lark in structure and materials, although it song flights are recorded in small numbers is generally more hidden under vegetation. outside the breeding season. Paired males spend Most nests found to date have been built on the much of their time guarding their mates, while ground under the mallow Abutilon pannosum unpaired males spend much of their time (most frequent in 2003) or under the Simple- singing and defending territories from other leaved Bean-caper Zygophyllum simplex (most unpaired males. One unpaired male in 2001 frequent in 2001), or occasionally in clumps of appeared to be defending two territories around grass [Poaceae]. De Naurois (in Bannerman & 500 m apart. Prior to copulation, males perform Bannerman 1968) recorded a number of nests hopping and bowing displays similar to those of built 10 cm or even 20 cm above the ground in the Sky Lark (Donald 2004). a ‘thick, low shrub’. Although he attributed Until recently, almost nothing was known these to the lark, it seems more likely they about the breeding ecology of the Raso Lark. belonged to the Cape Verde Sparrow Passer Breeding appears largely confined to periods iagoensis. Clutch size varies between one and Michael Brooke 212. Holes which Raso Larks Alauda razae have dug for food, Raso, Cape Verde, November 2003.

426 British Birds 99 • August 2006 • 420–430 The peculiar natural history of the Raso Lark three eggs (mean = 2.1), significantly fewer than quantify this, but already there are indications that recorded for the Sky Lark in any part of its that survival is indeed very high for a small bird range (Donald 2004). The low clutch size of living in such an austere environment. Raso Lark seems to fit the tendency noted by Bourne (1955) for many species in the Cape Feeding Verde Islands to have smaller clutches than their The Raso Lark depends for its existence on the mainland congeners. The eggs of Raso Lark are ability to find sufficient food on its arid island said by those familiar with both to closely home. Raso is small, has little vegetation and resemble those of the Wood Lark Lullula suffers periodic droughts, and it seems extraor- arborea, but are also very similar to, though dinary that birds can find enough food to noticeably smaller than, those of the Sky Lark. survive. That they do persist suggests a high As with most other larks, incubation is under- degree of adaptation, and an ability to exploit a taken entirely by the female (contrary to the range of different foods; it is only now that the account in BWP), and appears to last 12–13 remarkable range of adaptations shown by this days, a day or two longer than that of the Sky species to life on a desert island is becoming Lark (Donald 2004) despite the smaller clutch apparent. size. The female incubates in periods of around One of the most striking differences between ten minutes, followed by feeding breaks of the Raso Lark and its congeners is the structure around the same length, during which the male of its bill. Although only around two-thirds the often accompanies her. Although data are scant, size of a Sky Lark, the male Raso Lark has a it appears that the young may remain in the considerably longer and heavier bill. The bill of nest for up to 15 days after hatching, again con- the female is slightly larger than that of the Sky siderably longer than for the Sky Lark. Lark, but significantly smaller than that of the It is during the incubation period that an male. Indeed, unlike the Sky Lark, there is no extraordinarily high proportion of nests fail overlap in bill length of male and female Raso through predation. In October 2001 and again Larks. The likely explanation for the longer bill in November 2003, almost all nests were pre- of the Raso Lark is that this represents an adap- dated; in 2001, the overall survival rate of eggs tation to survival in a sandy habitat. Many from laying to hatching was estimated at just desert-dwelling larks have long bills for digging 5%. Because so many of the nests have been to uncover food buried in sand; an extreme predated by the time of hatching, the survival example of this adaptation is shown by the rates of chicks cannot be quantified, though we Hoopoe Lark Alaemon alaudipes of North have observed one case of chick predation in Africa and the Middle East. The significant dif- the nest. Almost all predated nests showed the ference in bill length between male and female same pattern of an undisturbed nest lining with Raso Larks is more difficult to explain, but large fragments of eggshells left in the nest, sug- recent observations of feeding behaviour may gesting that a single predator was responsible. provide a clue. Several early visitors to Raso This is almost certain to be the nocturnal Cape reported seeing holes in the ground which they Verde Giant Gecko Tarentola gigas, itself a variously assumed to be collapsed seabird threatened species restricted to Raso and the burrows or lark nesting holes. In fact, these are smaller Branco (Donald et al. 2003). This feeding holes, which are mined almost exclu- species is widely distributed on Raso but sively by males to extract the bulbs of appears to show an affinity for seabird burrows, nutsedges, probably the most important being where it may spend the day and gather much of Cyperus bulbosus or C. cadamosti. In areas its food (Hazevoet 1995). Raso Larks, like Sky where these plants occur, the ground may be lit- Larks, compensate in part for high nest losses tered with old and active holes. Individual holes by rapid re-laying after nest failure. One pair may have a productive life of several days, and was even observed building a new nest the day be worked by a number of different birds. following the loss of their previous nest. It is Digging is done with the bill, and the longer bill clear, however, that with irregular breeding and of male Raso Larks might explain why males high nest predation, productivity is very low; spend more of their time, and extract more of consequently, given the species’ continued exis- their food, by digging in this way. The females tence, it seems that adults must be relatively dig less frequently and obtain most of their long-lived. Ongoing studies of ringed birds will food by picking seeds and invertebrates from

British Birds 99 • August 2006 • 420–430 427 The peculiar natural history of the Raso Lark the surface and from under stones. The differ- lected from the song perches of males showed ence in bill structure between male and female the presence in the diet not only of vegetable Raso Larks may therefore represent an adapta- matter (present in all samples) but also of Lepi- tion that reduces competition for food between doptera larvae, beetle (Coleoptera) larvae, snails the sexes in a hostile environment with few (probably marine gastropods), and a number of resources. Each bulb takes an average of three other invertebrate groups (Donald et al. 2003). minutes of hard digging to find and extract, and Although data are lacking, it seems likely that clearly represents a valuable food item. Some chicks in the nest are fed largely on inverte- dominant males defend a number of feeding brates, and perhaps the availability of inverte- burrows from other birds in productive areas, brates limits breeding to periods immediately and others keep watch and move in to start following rainfall. digging when the ‘owner’ is occupied in chasing There is no standing fresh water on Raso, so off another bird. Others watch digging birds the lark must meet its water requirements (particularly the smaller females) and klep- through its food. Birds are often seen breaking toparasitise them when a bulb is uncovered. A off and chewing parts of the succulent plant wide range of tactics is therefore employed to Zygophyllum simplex, possibly to extract water. try to maximise bulb intake. On a small number of occasions, we have seen The rate of bulb intake by females is lower, larks fly down to the wave platform and appar- and the total time spent feeding higher, than that ently drink seawater from the surface of rock of the males, and females spend less time pools. The importance of this behaviour is preening and in vigilance behaviour (Donald et unclear. al. unpubl. data). This suggests that females might be under more pressure to find food than Conservation the males. If this is the case, it might explain the The Raso Lark is classified by IUCN as Criti- male-biased sex ratio, as females may be more cally Endangered (following recommendations susceptible to starvation. Intensive, long-term by Ratcliffe et al. 1999 to upgrade its status research by Peter and Rosemary Grant on from Endangered), placing it in a group of Darwin’s finches Geospiza in the Galapagos has around 200 bird species around the world con- demonstrated that differences in bill length as sidered most likely to suffer global extinction in small as 0.5 mm can have a profound impact on the near future. The Raso Lark has one of the survival rates, particularly during extreme cli- smallest ranges and populations of any bird. matic events (see Weiner 1995 for a popular Population trends are poorly known, but it is account of this work). After one particularly clear that, particularly during the long periods severe drought, most of the smaller-billed birds of drought that have historically blighted the in the Grants’ study population died. Although whole Cape Verde archipelago, numbers have we do not yet know enough about the Raso Lark dropped to extremely low levels, perhaps even to make direct comparisons, it is possible that the into single figures. There is little evidence to male-dominated population arises after pro- suggest that the current population is on longed drought through the higher mortality of average lower than it has been in the past, but a the smaller-billed females, perhaps because of number of potential threats exist, some of their lower intake of Cyperus bulbs. Intriguingly, which are growing. Historical records, conversa- there is no evidence of an uneven sex ratio in the tions with visiting fishermen from neigh- specimens collected by Alexander and others at bouring islands and the finding of desiccated the end of the nineteenth century, a period of scats indicate that cats and at least one dog have average or above-average Sahel rainfall. A study been present on Raso in the recent past, though of colour-ringed birds is underway to determine populations have apparently not established whether females do indeed have a lower survival and no mammals are thought to be present on rate than males, and long-term monitoring may the island at the time of writing. Raso Lark assess whether the sex ratio changes in the remains have not been found in the cat and dog females’ favour as rainfall increases. The Raso droppings examined (these contained almost Lark clearly has much to contribute to the study nothing but seabird remains). At least one of island biogeography. Short-eared Owl Asio flammeus, a rare visitor to Bulbs are not the sole source of food for Cape Verde, has been present on the island in Raso Larks. An analysis of faecal samples col- recent years, though again examination of

428 British Birds 99 • August 2006 • 420–430 The peculiar natural history of the Raso Lark pellets collected does not indicate the presence worse in the future (Held et al. 2005). This of Raso Larks in its diet (they also contained would reduce the average population of the mostly the bones of seabirds). Neglected Kestrel Raso Lark still further and make its extinction a and Brown-necked Raven Corvus ruficollis both probability. breed in small numbers (one to three pairs of Cape Verde is being promoted as a major each) on Raso but have probably done so for tourist destination, and an increasing number many years, and these species hunt infrequently of unregulated visits to Raso by tourists over the areas occupied by the larks. At present, unaware of the fragility of the island’s therefore, there is no suggestion that predators ecosystem poses a major threat. Visitors to Raso other than the gecko are impacting on the Raso can take precautions to minimise their impact Lark population. Nevertheless, growing on the larks and other birds, and can even con- numbers of visits to the island, likely to rise tribute positively to research and conservation. further as tourist developments on neigh- Most important is to take precautions to bouring islands are completed, increase the like- prevent the accidental introduction of lihood of predators becoming established. It is mammals, through appropriate packaging of also possible that non-native plant species will transported food and other supplies. Accidental be introduced, with an unpredictable outcome. introduction of non-native species to oceanic The one predator that does impact on the islands can occur very easily; research on lark is the Cape Verde Giant Gecko, which may another Atlantic island suggests that one insect have been present on the island for longer than species becomes successfully established every the lark (Carranza et al. 2000), and certainly three or four visits made by humans (Gaston et since well before human colonisation of the al. 2003). neighbouring islands. How the lark survives the Visitors arriving on Raso will often massive predation of its nests by the gecko is encounter local fishermen from São Vicente and unclear, yet clearly the two species have been Santo Antão, who stay for several nights each capable of surviving together in the past. week on Raso. Some of these fishermen illegally However, if human influence has increased the harvest young seabirds for food in October and population of the gecko, for example by the November, and may therefore resent the pres- provision of food in the form of scraps left by ence of visitors and the lark they come to see. visitors, it could be that nest survival drops The co-operation and goodwill of these below the level needed to sustain the lark’s pop- islanders is crucial to keeping Raso clear of cats ulation. The future survival of the lark and the and other predators, and visitors are encouraged gecko are clearly linked, and changes in the to take small gifts for them where possible and populations of one could influence those of the to treat them with the courtesy and respect they other. While we now know enough about the deserve. Once on the island, it is extremely lark to begin to understand its ecology, we important that visitors take great care where know next to nothing about the gecko, and this they walk or camp. Areas where holes are being an area of research that must be tackled if we excavated are extremely important, and great are to understand fully the Raso Lark’s ecology. care must be taken to avoid standing on and col- Increases in drought length and frequency lapsing them. If birds are seen carrying food or represent a clear threat, and rainfall since 1970 behaving nervously, it is likely that a nest is has generally been below the long-term average nearby and the area should be avoided. While on (fig. 1, p. 000). How predicted anthropogenic the island, visitors are likely to encounter birds climate change will affect rainfall patterns in fitted with coloured plastic and metal leg-rings; Cape Verde is unclear. Some models predict an the authors would be most grateful to receive increase in rainfall on oceanic islands (e.g. details of any ring combinations recorded. Whetton et al. 1996), which may bring the Raso Finally, visitors spending more than a day on the Lark some respite, but many areas of Cape island may have the time to get some estimate of Verde are suffering from desertification, and the population and distribution of Raso Larks, some of the larger islands are now scattered and possibly of any introduced mammals, and with abandoned villages. The most recent such information would be gratefully received climate models suggest that the recorded drying by the authors. of the Sahel zone since 1970 is partly due to The Raso Lark has a smaller world popula- greenhouse gases, and that it is likely to get tion than such conservation causes célèbres as

British Birds 99 • August 2006 • 420–430 429 The peculiar natural history of the Raso Lark the Californian Condor Gymnogyps californi- — 1898b. Further notes on the ornithology of the Cape anus, the Seychelles Magpie-robin Copsychus Verde Islands. Ibis 4: 277–285. Bakker,T., & van Dijk, K. 1996. Cape Verde Trip Report. sechellarum and the Mauritius Kestrel Falco Unpublished report. punctatus, all of which owe their survival to Bannerman, D. A., & Bannerman,W. M. 1968. Birds of the dedicated and expensive conservation initia- Atlantic Islands.Vol. IV.A history of the birds of the Cape Verde Islands. Oliver & Boyd, Edinburgh. tives. At present, no specific conservation mea- Bourne,W. R. P.1955.The birds of the Cape Verde Islands. sures are in place for the Raso Lark, although Ibis 97: 508–556. the species has been protected under Cape Burton, P.J. K. 1971. Sexual size dimorphism in Alauda Verde law since 1955. Raso was designated a razae. Bull. BOC 91: 108–109. Carranza, S., Arnold, E. N., Mateo, J. A., & López–Jurado, L. F. National Park in 1990 and official permission is 2000. Long-distance colonization and radiation in required to visit, although this is rarely sought gekkonid lizards,Tarentola Reptilia: Gekkonidae., and the island enjoys nothing more than revealed by mitochondrial DNA sequences. Proc. Roy. Soc. Lond. B 267: 637–649. nominal protection. This lack of conservation Castell, P.1999. Notes on the breeding biology of Raso effort is probably due to a number of factors, Lark Alauda razae. Bull.African Bird Club 6: 103–106. including the remoteness of Raso and the diffi- Donald, P.F. 2004. The Skylark. Poyser, London. culty of working there; the lack of an empow- ––, de Ponte, M., Pitta Groz, M. J., & Taylor, R. 2003. Status, ecology, behaviour and conservation of Raso Lark ered conservation community on Cape Verde; Alauda razae. Bird Conservation International 13: 13–28. the absence of any simple solutions; a lack of —, Brooke, M. de L., Bolton, M. R.,Taylor, R.,Wells, C. E., appreciation of the species’ status; and (dare we Marlow,T., & Hille, S. M. 2005. Status of Raso Lark Alauda razae in 2003, with further notes on sex ratio, suggest it?) the Raso Lark’s less-than-spectac- behaviour and conservation. Bird Conservation ular physical appearance. Regular monitoring of International 15: 165–172. the species, regular checks that mammals have Gaston, K. J., Jones, A. G., Hänel, C., & Chown, S. L. 2003. Rates of species introduction to a remote oceanic not reached the island and the careful manage- island. Proc. Roy. Soc. Lond. B 270: 1091–1098. ment of visitors to the island are minimum Hall, B. P.1963.The status of Spizocorys razae Alexander. conservation requirements, but the best chance Bull. BOC 83: 133–134. of ensuring the species’ long-term survival may Hazevoet, C. J. 1989. Notes on behaviour and breeding of the Razo Lark Alauda razae. lie in the eradication of cats from neighbouring — 1995. The Birds of the Cape Verde Islands. BOU Checklist Santa Luzía and the establishment of a popula- No. 13, BOU,Tring. tion of Raso Larks there. Recent visits confirm Held, I. M., Delworth,T. L., Lu, J., Findell, K. L., & Knutson,T. R. 2005. Simulation of Sahel drought in the 20th and 21st that there is much apparently suitable habitat centuries. Proc. National Academy of Sciences 102: on this larger island (Donald et al. 2005). That 17891–17896. the Raso Lark has survived for so long appears Lobin,W. 1982. Beitrag zur kenntnis der Cyperaceae miraculous, and it clearly has a great deal more Phanerogamae, Monocotyledonae. der Kapverdischen Inseln. Cour. Forsch. Inst. Senckenberg 52: 265–276. to tell us about island biogeography. How much Meinertzhagen, R. 1951. Review of the Alaudidae. Proc. longer it will survive without careful and sensi- Zool. Soc. Lond. 121: 81–132. tive conservation action is unclear. Nørrevang, A., & den Hartog, J. C. 1984. Bird observations in the Cape Verde Islands 4–22 June 1982. Cour. Forsch. Acknowledgments Inst. Senckenberg 68: 107–134. Ratcliffe, N., Monteiro, L. R., & Hazevoet, C. J. 1999. Status For help with fieldwork on Raso and support in many of Raso Lark Alauda razae with notes on threats and other ways, we are extremely grateful to Roy Taylor, Sabine foraging behaviour. Bird Conservation International 9: Hille, Mark Bolton, Colin Wells, Marta de Ponte, Maria Pitta 43–46. Groz, Mark Mainwaring,Tim Marlow and Justin Welbergen. Thorpe, R. S., Leadbeater, D. L., & Pook, C. E. 2005. We are also grateful to Keith Barnes, Norman Ratcliffe, Molecular clocks and geological dates: cytochrome b of Tom Spencer, José Tavares and Cornelis Hazevoet for their Anolis extremus substantially contradicts dating of help and advice, and to the many Cape Verdeans who Barbados emergence. Molecular Ecology 14: 2087–2096. have helped in various ways. Todd Mitchell, of the Weiner, J. 1995. The Beak of the Finch: a story of evolution in University of Washington, provided the data on Sahel our time.Vintage Books, New York. rainfall. We would particularly like to thank Julian Francis Whetton, P.H., England, M. H., O’Farrell, S. P.,Walterson, for his generous support of our expeditions to Raso. I. G., & Pittock, A. B. 1996. Global comparison of the regional rainfall results of enhanced greenhouse References coupled and mixed layer ocean experiments: Alexander, B. 1898a. An ornithological expedition to the implications for climate scenario development. Climatic Cape Verde Islands. Ibis 4: 74–118. Change 33: 497–519. Dr Paul F. Donald, RSPB, The Lodge, Sandy, Bedfordshire SG19 2DL; e-mail: [email protected] Dr M. de L. Brooke, Department of Zoology, University of Cambridge, Downing Street, Cambridge CB2 3EJ; e-mail: [email protected]

430 British Birds 99 • August 2006 • 420–430