Natalus tumidirostris Miller, 1900

Tejedor, Adrian, 2011, Systematics Of Funnel-Eared Bats (Chiroptera: Natalidae), Bulletin of the American Museum of Natural History 2011 (353), pp. 1-140 : 79-96

Natalus tumidirostris Miller, 1900

Figure 35

Natalus tumidirostris Miller, 1900: 160 . Type locality ‘‘Hatto, Curaçao, Netherlands Antilles.’’

Phodotes tumidirostris: Miller, 1906: 85 . New combination.

Phodotes tumidirostris continentis Thomas, 1910: 513 . Type locality ‘‘San Esteban, Carabobo, Venezuela’’; holotype, BMNH 5.11/12.25.

Natalus tumidirostris tumidirostris (part): Goodwin, 1959: 11. New combination.

Natalus tumidirostris continentis (part): Goodwin, 1959: 11. New combination.

Natalus tumidirostris haymani: Goodwin, 1959: 12 . Type locality ‘‘ Mt. Tamana , Trinidad, Trinidad and Tobago’’; holotype, AMNH 176590.

Natalus stramineus tronchonii: Linares, 1971: 81 . Type locality ‘‘Cueva de los Gavilanes, Rio Guasare , Zulia, Venezuela’’; holotype, MBUCV 1578.

Natalus stramineus: Linares, 1998: 515 . Not Natalus stramineus Gray, 1838 .

HOLOTYPE: USNM 102106, adult male, skin in alcohol with skull removed, collected by L.J. Guthrie on 5 January 1900, in Hatto [5Hato] (locality 467 in appendix 1), Curacao, Netherlands Antilles. Both the skin and skull are in good condition but the skin color has faded due to long immersion in alcohol. DISTRIBUTION: Mainland South America north of the Amazon River in the countries of Colombia, Venezuela, Guyana, Suriname, and French Guiana , and also on the islands of Trinidad, Tobago, Margarita, Curaçao, and Bonaire. Although to date it has not been reported from Brazil it very likely occurs there given that three collection localities (two in Guyana and one in Suriname) lie less than 15 km from the border with that country (fig. 36).

DIAGNOSIS: Size medium (forearm length 35.0–42.0mm), medial margin of ear pinna deeply concave; lateral margin of ear pinna deeply notched; nostrils usually large, opening anteriorly; maxilla above molars convex and markedly inflated; palate ending caudally always rostral to optic canal; ventral pelage monocolored; dorsal pelage monocolored or, if bicolored, hair bases lighter than tips; hair at base of claws short and inconspicuous or long and thin, never forming tufts; caudal margins of maxilla in ventral view forming an acute angle with longitudinal axis of skull; basisphenoid pit shallow; postorbital region with sides widely diverging rostrally, in dorsal view; caudal margin of ascending ramus of mandible perpendicular to alveolar margin of dentary; I1 not visible in lateral view, being obscured by I2; mesostylar crest of M3 absent. A comparison of diagnostic characters between N. tumidirostris , and other species of Natalus is summarized in table 5.

DESCRIPTION: Size medium (forearm length 35.0–42.0 mm; greatest skull length 15.5–17.3 mm; weight 4.3–8.9 g); muzzle long and dorsoventrally flattened; nostrils often wide and nearly circular, rarely elliptical; opening anteriorly on shallow depression on margin of upper lip; upper lip thickened; lower lip markedly thickened and constricted along dorsal and ventral margin, with numerous transversal grooves; small, smooth central pad on dorsal margin of lower lip; natalid organ medium size and elliptical, extending from caudal base of rostrum to crown of head; ears medium sized (13.0– 16.4 mm); ear pinna funnel shaped but distally thin; pinna with markedly pointed tip; medial and lateral margins of pinna deeply concave; five to six small ear ridges along lateral margin of distal pinna; ventral region of ear pinna greatly expanded, covering the eye and tragus in lateral view; medial ear margin thin and flexible; tragus short, lanceolate, and twisted into helixlike structure; tibia (18.9–22.4 mm) slightly longer than half the length of the forearm; calcar very long and thin, occupying about half the length of the free edge of uropatagium; free margin of uropatagium with sparse fringe of thin hairs; wings relatively broad, with 3rd metacarpal (33.7–38.7 mm) slightly longer than 5th metacarpal (30.4–38.3 mm); wings attach to tibia above ankle; pelage dense and lax; hairs long (7–9 mm, dorsally; 6–7 mm, ventrally); pelage usually darker dorsally than ventrally; pelage color from almost white ventrally and very light brown dorsally to bright orange brown both ventrally and dorsally (pl. 1); dorsal hairs bicolored, with tips darker than bases; ventral hairs usually monocolored; dense mustachelike hair tufts along lateral margins of upper lip and across muzzle; mustache formed by dense, lax, irregularly arranged, and ventrally curved hairs; natalid organ covered with thin hairs; skull long and relatively broad with moderate rostral flexion; rostrum wide, with sulcus between nasals almost imperceptible; moderate rostral palatal emargination; maxilla inflated, obscuring molars in dorsal view; braincase inflated, rising abruptly from rostrum; sagittal crest moderately developed; postorbital constriction wide; maxillary branch of zygomatic arch thin, less deep than twice the height of crowns of last molars; pterygoids nearly parallel; palate extending caudally to level of M2 or M3; basisphenoid pit shallow; longitudinal medial ridge on basisphenoid present; ectotympanic small, covering less than half of periotic; upper incisors short and peglike; I2 obscuring I 1 in lateral view; occlusal profile of premolars long; upper premolars of similar size; mesostylar crests on M1 and M2 short and straight, mesostylar crest absent on M3; cingular cusp of p4 medium sized and broad; molars cusps relatively broad; spinous process of humerus about as high as capitulum; thorax relatively short and wide; ribs markedly expanded craniocaudally and extensively in contact with one another; vertebrae C7 to T1 fused among themselves and to ribs; vertebrae T12–L4 fused entirely without vestige of sutures; vertebrae L5 and L6 free; caudal vertebrae 4 to 7 longer than distance from ischium to iliac crest of sacrum.

COMPARISONS: Fenestration of the palate is more extensive in N. tumidirostris than in any other species of Natalus . All palatal fenestrae, the interpterygoid and the intermolar, are very large and can coalesce with each other in various ways creating three main patterns: (1) all fenestrae separate, caudal edge of palate nearly at level with the caudal margin of M3 (common in individuals from Colombia); (2) intermolar fenestrae coalesced and separate from interpterygoid fenestrae, caudal edge of palate nearly at level with caudal margin of M3 (common in individuals from Sierra de Perijá); (3) all fenestrae coalesced forming deep palatal emarginations that reach caudal margin of M1 (common in individuals from the center and east of the species’ range). Occasionally, the degree of coalescence of the fenestrae is not symmetrical within a single individual, creating a combination of any two of the patterns mentioned above. In N. tumidirostris , thus, the caudal edge of the palate lies at the caudal margin of M1 or M3 whereas in all other species of Natalus the caudal margin of the palate lies at 1/2–2/3 of the way between the caudal margin of M3 and the tip of the pterygoid processes. Also, in all other species of Natalus , the intermolar palatal fenestrae, if present, are usually small and only rarely coalesce.

In Natalus tumidirostris , the maxillae are conspicuously inflated (although less so in individuals from Colombia and western Venezuela) and in dorsal aspect may hide the molars from view. In all other Natalus the labial margins of the molars are visible in dorsal view. In most specimens of N. tumidirostris , in lateral aspect, the rostral surface of the premaxilla forms a shallow (obtuse) angle with the dorsal surface of the nasals, whereas in all other species of Natalus this angle is less obtuse (even straight, in N. stramineus ). Also in lateral view, the dorsal surface of the nasals of N. tumidirostris is more parallel with the alveolar margin of the maxilla than in other species of Natalus , making the rostrum of this species appear relatively deeper anteriorly. In all other species of Natalus , especially in N. lanatus and N. primus , the rostrum tapers anteriorly more markedly.

Externally, Natalus tumidirostris can be distinguished from N. mexicanus , N. lanatus , N. primus , N. major , and N. jamaicensis by its concave medial margin of the ear. In N. tumidirostris , the medial margin of the ear pinna is almost as concave as the lateral margin, so that the ear tip appears acutely pointed. In the remaining aforementioned species, the medial margin of the ear pinna is straight to very slightly concave, so that the ear tip appears less acutely pointed. In most cases, N. tumidirostris can be distinguished from other species of the genus by the shape and size of its nostrils. The nostrils of N. tumidirostris tend to be as large as the external nasal septum, circular, and forwardly oriented. The nostrils of other species of Natalus are always smaller than the external nasal septum, elliptical, and usually oriented ventrolaterally. In some individuals of N. tumidirostris (as is common in Sierra de Perijá and on the island of Trinidad), however, these traits are less noticeable and the nostrils are similar in shape as in other species of the genus; thus, the shape of the nostrils should be used in combination with other traits, especially cranial ones, for a confident diagnosis of N. tumidirostris .

Natalus tumidirostris has a longer forearm and skull than N. mexicanus and N. lanatus . It is, however, smaller than the three Greater Antillean species N. primus, N. major , and N. jamaicensis . N. tumidirostris overlaps widely in size with two other species of the genus: N. espiritosantensis and N. stramineus . Nonetheless, its toothrow length is larger than that of N. espiritosantensis . From N. stramineus , N. tumidirostris is best identified on the basis of qualitative characters only.

VARIATION: Gómez-Laverde (1986) reported that young adults (individuals with fused epiphyses and silky grayish pelage) from Cueva de Macaregua, Colombia, had a mean zygomatic breadth significantly smaller than that of (older) adults with brownish or yellowish pelage.

Pelage color varies widely in Natalus tumidirostris . Some specimens from Cueva Ricardo Zuloaga, state of Miranda, Venezuela, are bright orange brown, whereas most individuals from Paraguaná peninsula are extremely pale, being pale buff dorsally and pale cream to white ventrally. The palest individuals from Paraguaná, in addition, show an extreme lack of skin melanin, having entirely light pink faces and ears. Several color variants are usually present in a single population, and it seems that the apparent lack of variants within a population is due primarily to small sample size.

Male Natalus tumidirostris are larger than females in 10 of 14 measurements (P, 0.05; table 22). Females from Colombia, however, had a longer forearm than males from Colombia, offsetting the species-wide sexual difference in this trait (interaction effect P, 0.05).

Natalus tumidirostris is the most geographically variable of the four continental species of Natalus . Four populations (Perijá, Paraguaná, Curaçao-Bonaire, and Trinidad) showed no overlap in a canonical variate

analysis based on one external (forearm length) and seven skull measurements. Specimens from Colombia, the coastal ranges of northern Venezuela, Venezuelan Guiana, and Trinidad are the largest in both external and internal measurements, whereas populations from Perijá, and Curaçao-Bonaire were the smallest. Specimens from Paraguaná were characterized by a relatively short forearm but a relatively large skull (P, 0.05, fig. 37).

Inflation of the maxilla is less marked in animals from the western part of the range ( Colombia, Perijá). Similarly, fenestration of the palate is less extensive in specimens from the western part of their range, especially in animals from Sierra de Perijá (pl. 7).

NOTE: The population from Perijá was originally described as a subspecies of N. stramineus (N. s. tronchonii) by Linares (1971) mainly on the basis of a supposedly unique pattern of palatal fenestration. Upon examining the skull of the holotype of N. s. tronchonii, I found that it had been incompletely cleaned and that the remaining soft tissue obscured a pattern of palatal fenestra- tion similar to that of specimens of N. tumidirostris from Colombia.

NATURAL HISTORY AND CONSERVATION: Natalus tumidirostris is known from 64 localities, including two represented by bone remains only. In 33 of the collection localities it has been found at day roosts, almost all being caves (30) and mines (2), except for a group of three individuals found roosting in a hollow rubber tree near Tamana caves ( Goodwin and Greenhall, 1961). Natalus tumidirostris has been found in caves varying from medium (e.g., Cueva Pararille; De Bellard Pietri, 1969) to large size (e.g., Ricardo Zuloaga, De Bellard Pietri, 1969), although it also probably occupies caves of smaller sizes. As in other species of Natalus , N. tumidirostris roosts in caves that are warm and humid, but avoids the warmest areas within the caves. In a study performed in Cueva del Guano and Cueva Piedra Honda (listed as ‘‘Cueva Barra Honda’’), Paraguaná, Venezuela, N. tumidirostris selected roost sites of about 30u C but moved to warmer (33u C) or cooler (28u C) spots when disturbed ( Genoud et al., 1990). Some of the caves inhabited by N. tumidirostris may contain high levels of ammonia (e.g., Cueva Ricardo Zuloaga). Most caves known to harbor N. tumidirostris are formed in limestone. Linares and Löbig-A. (1973) mentioned that N. tumidirostris shifts roosting caves throughout the year.

Roosts of Natalus are unknown in the Guyana Shield (southern Venezuela, the Guianas, and extreme northern Brazil), and all 12 records of the genus from this area correspond to captures with mistnets. In the Guyana shield, however, other cave-dwelling bats (i.e., mormoopids, Lonchorhina ) have been found in caves formed by the accumulation of large boulders around the eroding Precambrian granite inselbergs that dominate the Guianan landscape. It is therefore likely that on the Guianan shield Natalus tumidirostris also roosts in these types of caves.

Natalus tumidirostris has been found sharing roosts with 10 other bat species ( Pteronotus parnellii , Pteronotus gymnonotus , Pteronotus personatus , Mormoops megalophylla , Leptonycteris curasaoe , Phyllostomus hastatus , Carollia perspicillata , Artibeus jamaicensis , Desmodus rotundus , Myotis keaysi ) and roosts in closest association to Carollia perspicillata with which it often forms mixed groups, as in Macaregua ( Gómez-Laverde, 1986) and Cueva las Animas. Within roosting groups individuals hang widely spaced and are generally quiet, sometimes allowing themselves to be caught by hand. Occasionally, when large multispecies groups of bats abandon certain areas of a given cave due to the presence of visitors, individuals of N.

tumidirostris are the last to leave their roosting spots. Nonetheless, it has been observed to become more alert while roosting when a cave is repeatedly visited by humans ( Gómez-Laverde, 1986). It generally roosts in low areas of walls, but it is occasionally found hanging from high (3–5 m) ceilings.

Natalus tumidirostris has been found in habitats ranging from dry cactus scrub ( Bonaire, 464 mm annual precipitation; locality 465) to wet forest (Camp Patawa, French Guiana), but most commonly it is found in areas of deciduous to semideciduous forest. It occurs from sea level to middle elevations (e.g., Cueva Macaregua, Colombia, 1400 m; locality 67).

Natalus tumidirostris feeds on insects, and is reported by Linares (1998) to consume mostly Lepidoptera and Diptera . It seems to have a single annual estrus, and bears only one pup per year. The timing of the reproductive activity varies slightly among localities. In northern Venezuela, 13 (65 %) out of 20 females collected between 16–30 April 2004 were pregnant. In the only wellknown Colombian population, however, birth takes place before late March ( Gómez-Laverde, 1986). Throughout lactation, juveniles have been found to aggregate in compact clusters in cave ceilings during the daytime. Gómez-Laverde (1986) described the development during four months of one such groups in Cueva Macaregua, Colombia. In this cave, in late March, hairless neonatal individuals were found aggregated in a compact cluster on the cave ceiling, in association with a few adult females, which flew away after being disturbed. By mid- April, this group had consolidated into a large mass of many juveniles that occupied a 1 3 1.5 m of substrate. By the end of the month, these juveniles had appreciably grown, were already haired, and some were able to fly. In early June, the cluster had disappeared and, in its stead, there were many dispersed, adult-sized, and gray-colored individuals, all capable of full flight. The following year the cluster of juveniles was much reduced in size, reaching a maximum of only about 200 individuals, and was in a different location. It was not determined, however, whether this change reflected an overall smaller number of births or a change in the distribution of the newborns within the cave. In an incident in Macaregua, a piece of cave ceiling fell to the floor with about 10 nonflying pups still clinging to it; in spite of the accidental change in location, the pups continued to be fed, since observations made 15 days after the initial one revealed that they had grown appreciably and were still on the piece of fallen ceiling. A photograph taken in Cueva Ricardo Zuloaga, Venezuela, in an unspecified date, documents a similar clusterforming behavior of newborn N. tumidirostris ( Carreño, 1998) . Sex ratios appear widely skewed in some caves (e.g., an all male colony in Cueva del Guano in April, 2004) suggesting a pattern of sexual segregation similar to that of other natalids.

Natalus tumidirostris is an abundant species known from at least 961 museum specimens, 74 % of which having been collected in just three localities: Cueva Macaregua, Colombia; Cueva Ricardo Zuloaga, Venezuela; and Tamana Caves, Trinidad. It appears most densely distributed along northernVenezuela. Natural predators are not known, but at least in Cueva del Guano, Paraguaná, they could be occasional victims of giant centipedes ( Scolopendra gigantea , which is reported to actively hunt bats in this cave; Molinari et al., 2005). One of the few known specimens from French Guiana, was caught by a domestic cat that brought it into a house ( Charles-Dominique et al., 2001). Parasites of this species include Tricholeiperia trinidadensis (Nematoda, Molineidae ; Gibbons and Omah-Maharaj, 1991) and the bacterium Borrelia ( Marinkelle and Grosse, 1968) .

Harp trap surveys in Cueva Macaregua ( Cadena, 1974) indicate that nightly foraging begins relatively early at around sunset. The same study reported continuous activity of bats leaving and entering the cave until 22:00, when observations were stopped. Occasionally there was a peak in the number bats leaving the cave about half an hour after sunset.

The flight of Natalus tumidirostris is slow and very maneuverable, as in other species of the genus. If on the ground, N. tumidirostris seems unable to crawl, but can initiate flight vertically with strong downward thrusts of the wings ( Riskin et al., 2005).

As in other species of natalids, Natalus tumidirostris dies quickly (in fewer than 20 hr) of starvation and/or dehydration when kept in captivity. Its basal metabolic rate (1.54 ml O2/ghr) is very low (30 % below expected value) relative to that of other Neotropical insectivorous bats of similar body mass. Such a low basal metabolic rate may help this bat reduce its risk of starvation and water loss when roosting in warm caves and when foraging in dry habitats ( Genoud et al., 1990). Natalus tumidirostris seems able to survive lower temperatures than other small Neotropical bats due to its relatively low thermal conductance (0.41 ml O2/ghr uC), which is probably afforded by its long and lax hair ( Genoud et al., 1990). The thermoneutral zone of N. tumidirostris was determined to lie between 28u and 35u C, and this bat is able to maintain a stable body temperature of about 32u C within ambient temperature values ranging between 28u and 20u C, falling into torpor when the ambient temperature decreases below 20u C. It seems, however, unable to survive ambient temperatures below 10u C for longer than 2 hours ( Genoud et al., 1990).

Natalus tumidirostris is a species of least concern in IUCN’s Red List of Threatened Species (IUCN, 20102010). In northern Venezuela it seems a ubiquitous and locally abundant bat and therefore not threatened. Petit (1996) considered it to be threatened in Curaçao, with an islandwide population count of just 50–60 bats. The colony of Hato Cave, the source of the holotype of the species, has disappeared and the cave has been turned into a tourist attraction. Whitout appropriate management, this unique island population might become extinct.

ECOMORPHOLOGICAL DIVERSITY

Until recently, traditional taxonomists (with the notable exception of G.S. Miller) had failed to detect the subtle yet significant morphological variation that is present among members of the family Natalidae . As a consequence, intriguing patterns of morphological diversity within this group have been overlooked. Armed with a deeper understanding of the diversity of the family, these patterns can now be explored. Here, I examine morphology in an ecological context, and discuss their possible functional value.

Natalids show a range of body sizes (2 g in Nyctiellus lepidus to 12 g in Natalus primus ) and variations in morphology that are remarkable for a bat family of relatively low diversity (4 genera, 13 species). The anatomical modifications of some natalids represent extremes of morphological differentiation among New World bats, including the great extension of the flight membranes and the acquisition of accessory molar crests. In addition, one natalid genus, Chilonatalus , shows one of the greatest ranges in penis length within Chiroptera , as well as considerable variation in the size of the natalid organ, a glandular cephalic structure unique to Natalidae . These morphological phenomena have rarely been described, much less discussed in light of the ecology of the group.

In this section, available information on natalid ecology is used to address questions on the morphological diversification of the family. Three main questions are addressed in this section: (1) do different natalid species correspond to different ecomorphs of flight and feeding morphology? (2) is the structure of natalid faunas influenced by competition? (3) is genital morphology sexually selected in Natalidae ?

FLIGHT MORPHOLOGY: The relative size and shape (aerodynamic design) of flight surfaces (wings plus uropatagium) is highly variable among bats and has a profound influence on their ecology. Bats with pointed wing tips and narrow flight surfaces (high aspect ratio, defined as the ratio of the length to the width of a wing) fly fast, have little maneuverability, and exploit food resources in open spaces. Bats with broad wing tips and broad flight surfaces (low aspect ratio) tend to fly more slowly, have high maneuverability, and exploit food resources in cluttered habitats (i.e., within vegetation, McKenzie et al, 1995; Norberg and Rayner, 1987; Norberg, 1998; Stockwell, 2001). The maneuverability and flight speed of bats is significantly influenced by the size of the uropatagium, which represents a caudal extension of the flight membranes that is used in steering and adds aerodynamic drag. Bats with a large uropatagium tend to have higher flight maneuverability and to fly more slowly than bats with a small uropatagium ( Lawlor, 1973; Norberg, 1995).

Because of their wide wings and large uropatagia, natalids represent an extreme strategy among bats for slow, maneuverable flight ( Jennings et al., 2004; Norberg, 1998). Still, even within Natalidae there is significant variation in two measures of flightsurface design, the shape of the wing tip and the size of the uropatagium. With regard to wing-tip shape, natalids seem to comprise three distinct groups: (1) bats with relatively narrow wing tips ( Nyctiellus , Chilonatalus , and Natalus primus ); (2) bats with broad wing tips ( Natalus lanatus ); and (3) bats with wing tips of intermediate width (remaining members of genus Natalus ; figs. 38, 39). With regard to size of the uropatagium, natalids also fall into three categories: (1) bats with small uropatagia

( Nyctiellus ); (2) bats with large uropatagia ( Natalus , Chilonatalus tumidifrons , and C. macer ); and (3) bats with intermediate-sized uropatagia ( Chilonatalus micropus , Natalus lanatus ; figs. 38, 39).

Principles of bat-wing aerodynamics predict that natalids with narrow wing tips and small uropatagia will fly faster and with less maneuverability than natalids with broad wings and large uropatagia. Although natural history information that might corroborate these predictions is scant and vague, field observations seem to make a distinction between the flight of most natalids and that of Nyctiellus lepidus . With the exception of Nyctiellus lepidus , the flight of most species of Natalidae has been described as slow, very maneuverable, and even mothlike ( Goodwin, 1970; Jennings et al., 2004; Miller, 1905; Tejedor et al., 2004, 2005; the term ‘‘agile’’ given by some authors [e.g., Mitchell, 1965] is understood here to mean maneuverable). The flight of Nyctiellus lepidus , on the other hand, is relatively fast ( Silva-Taboada, 1979).

Despite the difficulty of interpreting the flight morphology of natalids in the absence of behavioral data, the differences found among genera and species are notable and suggest the existence of five distinct ecomorphs. Sympatric species normally belong to separate ecomorphs ( table 24 ). In this regard, the outlying position of N. lanatus is remarkable given that its sympatry with N. mexicanus represents the only known case of coexistence of two natalid species of the same genus. In only one case ( Chilonatalus macer and Natalus primus ) is the same ecomorph shared by two sympatric species.

MOLARIFORM DENTITION: Among bats, natalids show a particularly marked rostrocaudal elongation of the mesostyle of the upper molars. These modified molar cusps have been termed mesostylar crests by Morgan and Czaplewski (2003: fig. 4; pl. 16). Bat genera from other families, including Thyroptera (Thyropteridae) , Furipterus (Furipteridae) , and Kerivoula (Vespertilionidae) also show mesostylar crests, usually with a slightly different orientation and a lesser degree of development. Within Natalidae , the development of mesostylar crests is greatest in Primonatalus , Chilonatalus , and Natalus primus . In these taxa, the mesostylar crests are high, long, and broadly curved, occupying more than a third of the labial side of the occlusal surface of the upper molars. In Nyctiellus and the remaining species of the genus Natalus , the mesostylar crests are straight and relatively short (occupying less than 1/3 of the labial side of the occlusal surface of the upper molars). In addition, most members of Natalus (except N. primus ) lack a mesostylar crest on M3.

Long, sharp edges on molar crests appear within several predatory mammalian lineages (e.g., carnivorans: Butler, 1946; insectivorans and microchiropterans: Strait, 1993). The dilambdodont tooth of insectivorous bats, bearing multiple shearing crests on the Wshaped ectoloph, is a classic example ( Koopman and MacIntyre, 1980). These structures are thought to be optimal for shearing relatively soft food items as opposed to grinding coarser food, and are more highly developed in animals that specialize in eating boneless muscle ( Evans and Sanson, 2003) and/or soft-bodied insects ( Strait, 1993). Among insectivorous bats, the consumption of soft food items is also directly reflected in the shape of the skull. Bats with short, broad rostra can deliver stronger bites and are thought to specialize on eating hard-shelled insects (e.g., beetles), whereas bats with long, narrow rostra deliver weaker bites and consume soft-bodied insects (e.g., moths; Freeman, 1979).

The mesostylar crests of natalids represent additional shearing crests to the already welldeveloped dilambdodont teeth of these bats. Also, natalids show the greatest elongation of the rostrum among all insectivorous bats. This combination of traits suggests that natalids select soft food items. Moreover, the variation observed within the family in the development of the mesostylar crests seems to be associated with variation in the elongation of the rostrum.

Two main natalid groups are defined by rostral geometry: (1) bats with longer, narrower rostra ( Chilonatalus and Natalus primus ); and (2) bats with shorter, wider rostra ( Nyctiellus and the remaining species of Natalus ; fig. 40). Only species with long, narrow rostra show long and broadly curved mesostylar crests. Although the potential functional association of these two traits seems highly influenced by phylogeny (all species with long rostra and more shearing teeth descend from relatively basal nodes), it does indicate the existence of different natalid ecomorphs with relation to skull shape and molariform dentition. This variation probably reflects significant trophic niche differentiation among natalids.

Dietary data that could test this hypothesis is scant but hints at niche partitioning within Natalidae . The most complete dietary study of a natalid, Nyctiellus lepidus (a broadskulled species) reported a preponderance of insects of medium hardness (Homoptera, Diptera , and Hymenoptera ; Silva-Taboada, 1979). By contrast, accounts of the diet of narrow-skulled species mention a marked predominance of moths ( Chilonatalus macer ; Silva-Taboada, 1979) or moths and small crickets ( Natalus primus ; Tejedor et al., 2004).

As in the design of flight morphology, the putative ecomorphs of the feeding apparatus are partitioned among sympatric species with the exception of Natalus primus and Chilonatalus macer , which share the long rostrum/ long mesostylar crest ecomorph ( table 24 ). A second sympatric species pair, N. mexicanus and N. lanatus , also fall within the same ecomorph.

BODY SIZE: For a small family, Natalidae has a wide variation of body sizes. The largest member of the family, Natalus primus , is five times heavier and has almost twice the forearm length of Nyctiellus lepidus , the smallest natalid and one of the smallest bats in the world. Body sizes in Natalidae have a strong taxonomic component, with medium to large body sizes occurring only in the genus Natalus and small body sizes in Chilonatalus and Nyctiellus .

The distribution of body size in Natalidae appears to be related to sympatry among species (fig. 41). In three of the five natalid faunas with sympatric species there is a wide difference between the largest and smallest member of the assemblage. At present, the fauna of the Bahamas is unique within the West Indies in that it is composed of allopatric taxa ( N. lepidus and C. tumidifrons occur on different islands) and lacks a large species. During the last Pleistocene glaciation, however, the fauna of the Bahamas was similar to that of Cuba, having three sympatric species, and including an even larger-bodied representative of Cuba’s recent Natalus primus ( Morgan, 1989, 2001).

Differences in body size among closely related insular species are traditionally explained by two nonexclusive mechanisms: species assortment and/or character displacement. Species assortment states that only species with divergent morphology can successfully colonize and coexist on an island ( Grant and Abbot, 1980), whereas character displacement is defined as the divergence in morphology between species after the onset of sympatry as a result of selection due to competition ( Brown and Wilson, 1956). Rigorous demonstration of either mechanism requires proof of competition for resources and a genetic (rather than environmental) base for the observed morphological variation ( Schluter, 2000). Nonetheless, preliminary explanations seek support in the pattern of distribution and phylogenetic history of the taxa in question. The characteristic evidence for species assortment is that members of each size class are closest relatives. That for character displacement is that taxa that occur both allopatrically and sympatrically are more divergent between themselves when they occur sympatrically than when they occur allopatrically ( Losos, 1990).

Both patterns are present in Natalidae . The consistent generic difference in body size between Natalus (always large) and Chilonatalus (always small) in all islands where small and large natalids are found sympatrically is compatible with species assortment. By contrast, a unidirectional change in body size of the large ( Natalus ) and small ( Chilonatalus micropus ) natalids on the islands of Hispaniola, Jamaica, San Andrés, and Providencia, suggests character displacement. In Jamaica, where the largest Natalus ( N. jamaicensis ) of those islands occurs, C. micropus reaches its maximal body size, whereas in Hispaniola, where the smaller N. major occurs, C. micropus is also smaller (fig. 41). In San Andrés and Providencia, moreover, a member of the Antillean fauna ( Chilonatalus ) coexists with a medium-sized member of the continental fauna ( N. mexicanus ). In these two islands, N. mexicanus appears to have reached its largest body size and C. micropus its smallest, in agreement with the predictions of character displacement. Although the size difference between C. micropus and N. mexicanus on San Andrés and Providencia is much smaller than that between C. micropus and greater Antillean Natalus , the lack of differentiation in discrete characters of the insular N. mexicanus relative to that of the mainland (they are recognized as a single species) suggests that interaction between these two taxa is relatively recent and that the difference observed, if due to size divergence, may be only incipient.

Independently of the preponderance of one or the other mechanism in structuring natalid assemblages, two extreme cases, one of exaggerated size difference (between N. primus and C. macer in Cuba) and a second of lack of difference (between N. mexicanus

and N. lanatus in Mexico) suggest that competition and the evolution of body size are related in Natalidae . The large difference in forearm size between N. primus and C. macer may be related to the fact that these species share a similar morphospace. Natalus primus is unique in its genus in that it resembles Chilonatalus in its relatively long wing tip, large uropatagium, long rostrum, and high development of mesostylar crests of the upper molars. It is possible that the greater morphological similarity between N. primus and Chilonatalus has forced N. primus to become disproportionately larger to avoid competition. If body size is taken into account as an additional ecomorphological component, the pair N. primus / C. macer splits into clearly distinct groups in both wing shape and skull shape morphospaces (fig. 42 A, B).

The pair N. lanatus and N. mexicanus represent an opposite case. These two species are very similar in forearm length but have different flight morphology. Also, even though they both belong to the relatively wide-rostrum/short mesostylar crest ecomorph, they fall in opposite extremes of that ecomorph in the cranial shape ordination (fig. 42B). Separation into different flight and food-processing ecomorphs may thus be sufficient to partition resources between these two species, relaxing competition and the selective pressure to diverge in size. Nonetheless, as suggested before, it is possible that these two species may show some habitat segregation both in altitude and in roost selection, and that competition in this species pair may actually be minimal in nature.

GENITAL MORPHOLOGY: The relative length of the penis (length of the penis/ forearm length) is markedly variable among natalids. Three classes of penis length relative to body size can be distinguished in Natalidae : (1) penis long ( C. micropus ); (2) penis intermediate ( Natalus and Nyctiellus ); and (3) penis short ( C. macer and C. tumidifrons ; fig. 43). The mean relative penis length of C. tumidifrons (5.4 % of forearm length) is more than 10 standard deviations shorter than that of C. micropus (14.6 % of forearm length; figs. 43, 44).

Genital morphology has been shown to be associated with mating strategies and with the likelihood that females will mate with multiple males (Hosken and Stockely, 2004). In species with promiscuous females, sperm competition is intense and males are selected for production of large amounts of sperm (i.e., larger testes), for a deeper delivery of the ejaculate (i.e., longer penises), and for prolonged intromissions, all of which will increase the probability of inseminating females ( Arnqvist, 1998). Trends like these have been documented in primates ( Dixson, 1987; Harcourt et al., 1981; Verrell, 1992), carnivores ( Dixson, 1995), birds ( Briskie and Montgomery, 1997; Møller, 1988), flies ( Hosken and Ward, 2001), and butterflies ( Gage, 1994).

The positive association between intensity of sperm competition and testis size found in a wide range of taxa has been confirmed in bats ( Hosken, 1997, 1998; Wilkinson and McCracken, 2003). Investigations of the relationship of penis morphology to mating system in bats, however, have been inconclusive. In a study including 163 species of 12 bat families, significant positive correlations were found among relative baculum length, relative testis size, and mating system (Hosken et al., 2001). The relationships did not hold, however, after correction for phylogenetic history. The authors, therefore, speculated that other factors unrelated to sexual selection could influence baculum length, including aspects of female tract morphology and the size of the uropatagium, the last seen as a physical obstacle during copulation.

Natalids are an interesting group in which to explore genital evolution in bats given their significant variation in penis length and relative size of the uropatagium. Comparisons of penis length with length of the tibia (a correlate of size of the uropatagium in Natalidae , see Methods) showed no familywide correlation (fig. 45). Stronger trends, however, were detected within two species groups that differ in the relative size of the natalid organ: (1) bats with large natalid organs, reaching 13 % –26 % the length of the forearm; and (2) bats with small natalid organs, reaching 6 % –16 %. Interestingly, across Natalidae , relative penis length does appear to be negatively associated with relative length of the natalid organ (fig. 46).

The natalid organ, a unique synapomorphy of Natalidae ( Simmons, 1998) , is a presumed exocrine gland that is present in the forehead of adult males only ( Dalquest, 1950; Goodwin, 1959). The size and shape of the natalid organ varies widely among natalid genera (fig. 47; pl. 5), reaching its maximum development (8.5 mm; 53 % of the skull length) in C. tumidifrons . Although the function of the natalid organ is unknown, its exclusive presence in males suggests it has a sexual function. When manipulated, living males of the genus Natalus sometimes secrete a drop of an oily, translucent green liquid through a pore on the anterodorsal surface of the natalid organ. This secretion does not dissolve in alcohol, as individuals with hardened amber-colored droplets are occasionally found among specimens preserved in fluid. It is possible that secretions from the natalid organ function as a social signal in male-female and/or male-male interactions. The apparent negative association of the size of this organ with male genital size strengthens the hypothesis that both penis length and natalid organ are sexually selected in natalids.

The presence of opposite evolutionary trends in genital size within Chilonatalus is remarkable, and suggests significant differences in mating system among species of this genus. The long penis of C. micropus might be related to high levels of sperm competition in this species, perhaps entailed by high levels of female promiscuity. Conversely, the extremely short penis of C. tumidifrons and C. macer , together with the extreme development of their natalid organ suggest a different strategy. In these two species sperm competition could be reduced by an influence of this gland in social communication. It is possible that individuals with some extreme quality of the natalid organ or its production may have more exclusive access to females, as may occur among harem-keeping or lekking species with exaggerated secondary sexual dimorphism ( Andersson, 1994). These intriguing hypotheses remain to be tested with behavioral data.

Tables

TABLE 22 Summary of measurementsa of Natalus tumidirostris

		Females					Males
	N	Mean	(Min.–Max.)	SD	N	Mean	(Min.–Max.)	SD
Weight	29	6.4	(4.5–8.9)	1.1	28	6.3	(4.3–8.6)	1.0
Forearm length*	94	38.7	(36.1–41.5)	1.1	146	38.8	(35.0–42.0)	1.4
Length of tibia, dry*	25	18.9	(17.3–20.0)	0.7	28	19.7	(17.6–22.0)	1.1
Length of tibia*	34	20.3	(18.9–21.9)	0.8	59	20.8	(18.9–22.4)	0.8
Length of 3rd metacarpal	23	36.2	(34.5–37.9)	1.1	39	36.2	(33.7–38.7)	1.2
Length of 5th metacarpal	23	35.3	(30.4–37.2)	1.4	39	35.5	(33.2–38.3)	1.3
Length of ear	57	14.9	(13.0–17.3)	1.1	79	14.9	(13.0–17.2)	0.9
Length of penis		–	– –	–	34	3.5	(2.4–4.9)	0.6
Length of natalid organ		–	– –	–		–	– –	–
Greatest skull length*	67	16.5	(15.5–17.2)	0.3	105	16.7	(15.5–17.3)	0.4
Zygomatic breadth*	66	8.3	(7.7–8.8)	0.2	101	8.4	(7.7–9.0)	0.2
Braincase breadth*	69	8.0	(7.4–8.4)	0.2	103	8.1	(7.3–8.7)	0.2
Breadth across molars*	71	5.5	(5.0–6.0)	0.2	105	5.5	(5.1–5.9)	0.2
Breadth across canines*	69	3.8	(3.4–4.1)	0.2	101	3.9	(3.3–4.2)	0.2
Maxillary tooth row *	73	7.0	(6.5–7.2)	0.2	103	7.1	(6.7–7.4)	0.2
Mandibular tooth row*	63	7.4	(6.8–7.7)	0.2	89	7.5	(6.9–7.8)	0.2
Postorbital breadth	66	3.4	(3.2–3.6)	0.1	106	3.4	(3.0–4.0)	0.2
Depth of braincase*	36	6.5	(6.1–7.2)	0.2	62	6.7	(6.2–7.2)	0.2

a Descriptive statistics of measurements for each sex. N 5 sample size; SD 5 standard deviation. Weight is given in g; all other measurements are given in mm. See text for description of measurement methods. Measurements significantly different between sexes (P, 0.05) are followed by an asterisk (*).

TABLE 24 Distribution and ecomorphs of Natalidae

		Flight ecomorph		Feeding ecomorph
		Shape of wing	Size of		Mesostylar	Body-size
Distribution	Taxon	tip	uropatagium	Skull	crests	ecomorph
Florida	Ppr	—	—	—	long	—
Bahamas	Nyl	pointed	small	broad	short	small
	Ctu	pointed	large	narrow	long	small
Cuba	Nyl	pointed	small	broad	short	small
	Cma	pointed	large	narrow	long	small
	Npr	pointed	large	narrow	long	large
Jamaica	Cmi	pointed	intermediate	narrow	long	small
	Nja	intermediate	large	broad	short	large
Hispaniola	Cmi	pointed	intermediate	narrow	long	small
	Nma	intermediate	large	broad	short	large
Lesser Antilles	Nst	intermediate	large	broad	short	intermediate
Mexico and	Nme	intermediate	large	broad	short	intermediate
Central America	Nla	rounded	intermediate	broad	short	intermediate
South America	Ntu	intermediate	large	broad	short	intermediate
north of
Amazon River
South America	Nes	intermediate	large	broad	short	intermediate
south of
Amazon river