U·M·I University Microfilms International a Bell & Howell Information Company 300 North Zeeb Road
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copysubmitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sectionswith small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
U·M·I University Microfilms International A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. M148106-1346 USA 313:761-4700 800/521·0600
Order Number 9325033
Character release in the endangered Hawai'ian hoary bat, Lasiurus einereus semotus
Jacobs, David Steve, Ph.D.
University of Hawaii, 1993
Copyright @1993 by Jacobs, David Steve. All rights reserved.
U·M-I 300 N. ZeebRd. Ann Arbor, MI48106
CHARACTER RELEASE IN THE ENDANGERED HAWAI'IAN HOARY BAT,
LASlURUS CINEREUS SEMOTUS.
A DISSERTATION SUBMITIED TO THE GRADUATE DMSION OF THE UNIVERSllY OF HAWAI'IIN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
IN
ZOOLOGY
MAY 1993
By
David S. Jacobs
Dissertation Committee:
Leonard Freed, Chairperson Sheila Conant Michael Hadfield Stephen Palumbi Rebecca Cann © Copyright 1993 by David S. Jacobs
iii To my Family and the Village that raised me.
iv ACKNOWLEDGEMENTS
I am indebted to F. Howarth and A Ziegler for providing unpublished manuscripts and information on work in progress, to C. Koehler and M. Kalcounis for providing wing tracings of the North American hoary bat, and to the Bernice
P. Bishop Museum (Honolulu) and the Museum of Southwestern Biology (New
Mexico) for providing bat skulls. This work was supported by grants from the
Audubon Society, the state Division of Forestry and Wildlife, Hawai ' i, U. S. Fish and Wildlife, Hawai-Ian Telephone Co. and Alvin Y. Yoshinaga.
v ABSTRACT
The insectivorous Hawaiian hoary bat, Lasiurus cinereus semotus, is an endemic subspecies of the North American hoary bat, L. c. cinereus. This study investigated morphological and ecological divergence of the island population, with emphasis on the potential for character release.
Divergence in the Hawai \ian hoary bat involved characters related to flight and feeding. The Hawai \ ian bat has undergone a 45% reduction in body size with allometric responses in size of its wings. These changes decreased wing loading
(ratio of mass to wing area), without altering the high aspect ratio
(wingsparr'jwing area) of the ancestral species, thereby permitting slower and more maneuverable flight near vegetation. High aspect ratio enables the bat to retain the fast and efficient flight of its ancestor in open areas. This increased flexibility in flight behavior has allowed the Hawai\ ian bat to expand its foraging habitat to include the open habitats of the North American hoary bat and the closed habitats of putative competitors of the North American hoary bat.
There has been a relative increase in skull length of the Hawai \ ian bat, with non-allometric increases in the gape of the jaws, height of the coronoid process, and size of the masseter muscle. This gives the jaw more crushing power for more efficient processing of large and hard-bodied prey. In the open habitat the
Hawai \ ian bat fed predominantly on moths as does the North American hoary bat. In the closed habitat the Hawai \ ian bat fed predominantly on beetles as do putative competitors of the North American hoary bat. The morphological and vi ecological changes are consistent with the assumptions and prediction of the character release hypothesis. However, uncertainties remain about the historical aspects of the assumptions.
In addition, the interaction between flight speed, echolocation, and insect prey was investigated. Apparent prey selection by fast-flying bats in open habitats was the result of a decrease in detectability of smaller prey. Bats responded to changes in prey density by altering the pursuit component of handling time.
vii TABLE OF CONTENTS
Acknowledgements ~ Abstract. Vl List of Tables ~ ix List of Figures .x Chapter 1: Introduction 1 Chapter 2: Distribution and Abundance of the Endangered Hawai \ ian Hoary Bat, Lasiurus cinereus semotus, on the island of Hawai \i. 15 Introduction 15 Materials and Methods 16 Results 18 Discussion 24 Chapter 3: Morphological Divergence in the Hawai \ ian Hoary Bat, Lasiurus cinereus semotus 27 Introduction 27 Methods 32 Results 35 Discussion 41 Chapter 4: Foraging Strategies and Prey Selection in the Hawai \ ian Hoary Bat, Lasiurus cinereus semotus 49 Introduction 49 Methods 54 Results 58 Discussion 62 Chapter 5: Character Release in the Hawai\ ian Hoary Bat, Lasiurus cinereus semotus 68 Introduction 68 Methods 72 Results 73 Discussion 75 Appendix: Museum Skull Specimen Numbers and Standard Errors of Skull Measurements 125 Literature Cited 126
viii LIST OF TABLES
2.1. Maximum Number of Bats at Each Foraging Site 81
2.2. Temporal Distribution of Bats 82
2.3. Vegetation Associated with Each Foraging Site 83
3.1 Body Mass and Forearm Data 84
3.2. Skull Dimensions 85
3.3. Allometric Exponents of Skull Measurements 86
3.4. Wing Parameter Data Wingspan, Wing Area, Wing Loading 87
3.5. Allometric Exponents of Wing Parameters 88
3.6. Wing Parameter Data Aspect Ratio and Tip Shape Index 89
4.1. Size of Dietary Items 90
4.2. Percentage of Artificial Prey Chased 91
4.3. Bat Acoustic Responses to Artificial Prey 92
4.4. Percentage of Capture Attempts on Artificial Prey 93
A.l. Museum Skull Specimens 125
A2. Standard Errors for Repeated Measures l25
ix LIST OF FIGURES
Figure
2.1. Distribution of Bat Observations 94
2.2. Bat Activity and Insect Biomass 95
3.1. Wing Design for Closed and Open Habitats 96
3.2a. Skull Measurements 97
3.2b. Skull Measurements 98
3.2c. Skull Measurements 99
3.3. Definitions of Wing Parameters and Measurements Measurements 100
3.4. Plot of Mass against Skull Length 101
3.5. Plot of Mass Against Wingspan 102
3.6. Plot of Mass Against Wing Area 103
3.7. Plot of Mass Against Wing Loading 104
3.8. Plot of Mean Mass Against Mean Wing Loading 105
4.1a. Insect Orders in Bat Diet and Light Traps Closed Habitat (Ocean View), August 1991 106
4.1b. Insect Orders in Bat Diet and Light Traps Closed Habitat (Ocean View), September/October 1991.. 107
4.1c. Insect Orders in Bat Diet and Light Traps Closed Habitat (Ocean View), June 1992 108
4.1d. Insect Orders in Bat Diet and Light Traps Closed Habitat (Ocean View), August/September 1992 109
4.2. Insect Orders in Bat Diet and Light Traps Closed Habitat (Red Cinder Road) 110
x LIST OF FIGURES (contd.)
Figure
4.3a. Insect Orders in Bat Diet and Light Traps Open Habitat (Pohakuloa), June 1992 111
4.3b. Insect Orders in Bat Diet and Light Traps Open Habitat (Pohakuloa), July 1992 112
4.3c. Insect Orders in Bat Diet and Light Traps Open Habitat (Pohakuloa), September 1992 113
4.4a. Proportion of Bats Eating each Insect Order Closed Habitat (Ocean View), August, 1991.. 114
4.4b. Proportion of Bats Eating each Insect Order Closed Habitat (Ocean View), September/October 1991.. 115
4.4c. Proportion of Bats Eating each Insect Order Closed Habitat (Ocean View), June 1992 116
4.4d. Proportion of Bats Eating each Insect Order Closed Habitat (Ocean View), August, 1991.. 117
4.5. Proportion of Bats Eating each Insect Order Closed Habitat (Red Cinder Road), August/September 1992 118
4.6. Proportion of Bats Eating each Insect Order Open Habitat (Pohakuloa) 119
4.7. Size Distribution of Insect Orders Closed Habitat (Ocean View) 120
4.8. Size Distribution of Insect Orders Closed Habitat (Red Cinder Road) 121
4.9. Size Distribution of Insect Orders Open Habitat (Pohakuloa) 122
5.1. Size Range of Prey Items 123
5.2. Proportions of Moths and Beetles in the Bat Diet. 124 xi CHAPTER 1
INTRODUCTION
''/ should like to know whether the case ofendemic bats in islands struck you: it has me especially; perhaps too strongly" (Charles Darwin in a letter to J. D.
Hooker, dated March 15, 1859 -Burkhardt and Smith, 1991).
DIVERGENT EVOLUTION IN INSULAR POPULATIONS
Divergent evolution occurs when the descendants of an ancestral population develop different heritable phenotypes. Anagenesis is a type of divergent evolution that involves change within the same lineage over evolutionary time without speciation (Futuyma, 1986, pp.403). Such change may include geographical variation of a single character (e.g., beak depth) within the same species. Cladogenesis, on the other hand, involves change between an ancestral population and some of its genetic lines that become isolated. One population splits to give rise to two or more divergent populations that may become different species. Whereas anagenesis involves a gradual and directional change within a lineage, cladogenesis may involve rapid change as different gene pools encounter different environments. The distinction between anagenesis and cladogenesis does not imply that the two processes necessarily occur in isolation of each other.
Evolution may occur by anagenesis within a population that then undergoes cladogenesis, with anagenetic changes in each split lineage until the next cladogenetic event (Maurer et al., 1992).
1 Independent evolutionary trajectories associated with cladogenesis require reproductive isolation. Such isolation may arise within a population giving rise to sympatric divergence. One example is the development of tetraploidy by hybridization between two diploid plants (Futuyma, 1986; pp.228). Reproductive isolation leading to allopatric divergence may arise as a result of geographical barriers to gene flow (e.g., a mountain range, a river, a lava flow, or open ocean).
The most extreme form of reproductive isolation that leads to allopatric divergence involves founder events in which only a fraction of the ancestral gene pool is present in the new population (Mayr, 1954).
Divergence of oceanic island fauna and flora from their continental parent populations is perhaps the classic example of cladogenesis in allopatry involving a founder event in a new environment. The biota of islands are characterized by a high level of endemism (Darwin, 1950, pp.346; Foster, 1964; Zimmerman, 1970;
Sondaar, 1977), and many species have morphological and ecological characteristics that differ greatly from those of the ancestral population. A great deal of attention has been paid to the divergent evolution of oceanic island fauna from their parent populations, and the factors responsible for such divergence
(e.g. Darwin, 1950; Lack, 1947; Diamond, 1970; 1978; Grant, 1986).
Morphological divergence on islands has been associated with changes in locomotion. Flightlessness in insular bird species has evolved more than once, e.g., the kiwis of New Zealand, and the dodos on islands like Mauritius in the Indian
Ocean (Carlquist, 1965). The loss of legs in certain genera of skinks (e.g.,
2 Brachymeles) is another example of morphological divergence on islands
(Carlquist, 1965). Morphological changes have also been associated with profound changes in diet. Divergence in beak size and shape in Darwin's finches on the
Galapagos (Lack, 1947; Grant, 1986), and in the Hawai\ ian Honeycreepers
(Carlquist, 1974) has been accompanied by an impressive divergence in diet. From a single ancestor, some of Darwin's finches developed heavier bills to crack harder seeds, and others developed long narrow bills for probing into narrow cracks for insects (Grant, 1986). In Hawai \ i, some moth caterpillars have become carnivorous and some crane-fly larvae mine leaves (Zimmerman, 1970;
Montgomery, 1983).
Ecological divergence in habitat and elevation has occurred among bird species on islands like Puercos Island, in the Pearl archipelago (MacArthur et al.,
1972) and on the island of Kakar, off New Guinea (Diamond, 1970). The island of
Kakar has 49 species of birds, 11 of which have expanded their altitudinal range and 12 of which have expanded their habitat. The swift, Collocalia esculenta, has expanded its range to include forests where it hawks insects in much more confined spaces and makes much tighter turns than elsewhere (Diamond, 1970).
At least one species, the finch Lonchura tristissima, has also expanded its diet from seeds to include insects (Diamond, 1970).
This dissertation concerns the allopatric divergence of an island bat, the
Hawai \ ian hoary bat, Lasiurus cinereus semotus, from its continental parent population. This bat is thought to be derived from the North American hoary bat,
3 Lasiurus cinereus cinereus (Sanborn and Crespo, 1957;Tomich, 1986; Fullard,
1987), and is the only extant bat in Hawai' i. Previous studies suggest that divergence has occurred in the Hawai ' ian hoary bat: it has undergone a reduction in size (Tomich, 1986) and appears to have a more generalized diet than its North
American counterpart (Whitaker and Tomich, 1983). This study investigates morphological divergence in the Hawai'ian hoary bat and the relationship of such divergence to the foraging ecology of this bat.
INSULAR DWARFISM
Changes in body size, such as observed in the Hawai ' ian hoary bat, have most often been used to quantify the divergence of island fauna (e.g. Lack, 1947;
Foster, 1964; Diamond, 1978; Grant, 1986). Size change in insular mammals follows a tendency towards gigantism in rodents and marsupials, and towards dwarfism in carnivores, lagomorphs, and artiodactyla, as well as in fossils of island elephants, hippopotami, deer, and marsupials (Foster, 1964; Sondaar, 1977;
Marshall and Corruccini, 1978; Lister, 1989). These patterns of insular size change have been termed the "island rule" by Van Valen (1973). However, questions about the validity of the island rule have been raised. Lawlor (1982) suggests that lagomorphs and heteromyid rodents do not exhibit the predicted trends in insular body size. Lomolino (1985), re-examining the island rule, found that rather than describing differences among mammalian orders, the island
4 rule should account for the graded trend from dwarfism in the larger species of insular mammals to gigantism in the smaller species.
Special attention has been focused on insular dwarfism in mammals (e.g.
Foster, 1964; Gould, 1975; Sondaar, 1977; Marshall and Corruccini, 1978). A number of factors have been advanced as causes of the decrease in the size.
Darwin (1950, pp.298) attributed divergence more to "new relations" with organisms in the new environment than to the new set of physical conditions to which the founding population was exposed. Sondaar (1977) suggested that dwarfism in island mammals is an adaptation to an island environment in which large carnivores are absent. Smaller size would be advantageous in the resource and space-limited island environment in the absence of predation pressure. It allows greater mobility, decreases the quantity of food and territorial area required, and increases heat exchange with the surrounding air or water.
According to Bergmann's rule (Scholander, 1955; Smith, 1980), the larger body surface area to volume ratio of smaller animals facilitates thermoregulation in cases where the island population experiences a warmer climate. Similarly,
Marshall and Corruccini (1978) suggested that dwarfism is the result of selective elimination of large individuals by limitation of resources. The selection is expected to maintain allometry in the rest of the population.
Lomolino (1985) contended that competitive release (divergence as a result of the absence of interspecific competition) and resource limitation, rather than predation perse, appear to be the major factors affecting insular body size in
5 mammals. The importance of competitive release as a factor for promoting gigantism decreases with body size. This explanation assumes that larger individuals have greater niche breadths (i.e., can feed on large and small prey) so that in the absence of competitors in an insular environment a smaller organism can expand its niche breadth by becoming larger. The relative importance of competitive release decreases with size since larger species are less likely to be those whose body size is constrained by competition on continents. In contrast, the importance of resource limitation in promoting dwarfism increases with body size because larger individuals require greater resources. Limited food supplies would select for smaller individuals since they require less resources. These perspectives account for the general form of the island rule in which insular mammals exhibit trends ranging from gigantism in the smaller species to dwarfism in the larger species (Lomolino, 1985).
CHARACTER RELEASE
The absence of competition has been advanced as an explanation for many of these examples of insular divergence (e.g. Carlquist, 1965; Diamond, 1970;
Lack, 1976). Island fauna and flora are depauperate and disharmonic (Foster,
1964; Diamond, 1970; Sondaar, 1977). Some taxa are represented very well and others completely absent. This is the result of chance and the differential dispersal abilities of organisms making up the pool of potential colonizers from continents.
Some taxa are good swimmers, or good fliers, and are able, at least occasionally,
6 to cross the water barrier by flying, swimming, rafting, being carried by birds, or being blown off course from their migratory routes. Other taxa are unable to reach the oceanic islands at all. The result is an environment of changed interspecific interactions such as predation and competition. An organism previously exposed to interspecific competition on the continent may be exposed to new competitors, or have no competitors in its new insular environment.
Islands are therefore suitable sites for investigating the potential effects of release from interspecific competition on the morphology and ecology of an organism.
Ecological theory regards interspecific competition as a significant selective agent that shapes the morphology and behavior of organisms (Darwin, 1950; Lack,
1947; Brown and Wilson, 1956; Diamond, 1978; Connell, 1983; Grant and
Schluter, 1984; Grant, 1986). Competition theory was formalized in the
Lotka-Volterra equations. These equations, supported by experiments conducted by G. F. Gause, gave rise to the "competitive exclusion principle" which states that two similar species utilizing the same resources cannot co-exist; one necessarily displaces the other (Smith, 1980). Emphasis shifted away from the competitive exclusion principle when Brown and Wilson (1956) drew attention to a related evolutionary phenomenon which they called "character displacement".
They defined this as "the situation in which, when two species of animals overlap geographically, the differences between them are accentuated in the zone of sympatry and weakened or lost entirely in the parts of their range outside this zone" (i.e., in allopatry).
7 Grant (1972) separated the concepts of sympatric divergence and allopatric release by redefining the former as "the process by which a morphological character state of a species changes, under natural selection, arising from the presence (italics mine), in the same environment, of one or more species similar to it ecologically and/or reproductively". He called this phenomenon character displacement. Grant defined character release as the process by which a morphological character state of a species changes, under natural selection, arising from the absence, in the same environment, of one or more species similar to it ecologically and/or reproductively. The fundamental difference between these two phenomena appears to be whether the founding population was sympatric or allopatric with respect to a competitor. If the species was allopatric and then became sympatric in some regions of its range, one would be dealing with a case of character displacement. If, however, it was sympatric and then became allopatric, character release is the applicable concept. This study investigates character release in the foraging ecology of the Hawai \ian hoary bat.
Grant (1972) restricted the term character to morphological attributes of an organism and considered parallel ecological and behavioral phenomena (e.g., changes in foraging behavior, habitat, or diet) as examples of ecological release.
This will also be done in the present study. Although ecological release may occur without character release, the latter can only be recognized if it is associated with an expansion of the foraging niche of the animal (e.g., an expansion in the
8 foraging habitat and/or diet of the animal). Ecological release will here be considered as a means for identifying character release.
Character release has been demonstrated in a few island populations of birds (Fjeldsa, 1982; Grant, 1986). Ecological release on islands has been reported in a variety of organisms including birds (Grant, 1965; 1966; Crowell, 1962; Keast,
1968; Diamond, 1970; Fjeldsa, 1982; Grant, 1986), rodents (Cameron, 1964), and lizards (Williams, 1969). However, in most cases of ecological and character release the assumption that natural selection is responsible for the release has not been evaluated (Grant, 1972).
Darwin recognized the importance of bats to geographical variation in his major work On the Origins ofSpecies by Natural Selection (1950, p.334), as well as in a letter to J. D. Hooker in 1859 (see frontispiece). However, to my knowledge, no one has reported character or ecological release in island or mainland bats.
This is surprising since bats are the only terrestrial mammals (because of their flying abilities) to have reached isolated oceanic islands (e.g. Hawai> i and the
Galapagos) without human assistance. It is precisely on such isolated islands that the greatest divergence attributable to character release would be expected.
CHARACTER RELEASE AND DIVERGENCE IN THE HAWAI'IAN HOARY
BAT
This study investigated divergence of the Hawai ' ian hoary bat from its
North American counterpart, and evaluated character release as an explanation
9 for the divergence. Both subspecies are aerial insectivores, i.e., they fly continuously while foraging and hunt flying insect prey (Whitaker and Tomich,
1983; Barclay, 1985). The North American hoary bat occurs sympatrically with a number of other insectivorous bat species (Whitaker, 1972; Black, 1974; Barclay,
1985). It appears that partitioning of the habitat, spatially and temporally (Jones,
1965; Kunz, 1973), and insect prey, by order and size (Ross, 1967; Black, 1972;
1974; Warner, 1985), occurs in these North American bat communities.
Apart from one other species of bat found in the fossil record on the islands of Kaua'i, 0'ahu, Moloka \ i, and Maui (no fossils from the island of
Hawai \ i), the Hawaiian hoary bat is the only extant bat in Hawai \ i. It thus represents the allopatric condition of the North American hoary bat with respect to related competitors, and appears to have evolved with less direct interspecific competition. Whitaker and Tomich (1983) suggest that the more generalized diet of the Hawai ' ian bat is due to competitive release.
Character release in a bat can be manifested in its wing and skull morphology, as well as in its echolocation call characteristics. The echolocation call characteristics of the Hawai \ian hoary bat have been investigated (James
Fullard, David Jacobs, Robert Barclay, mans. in prep.) and although the main conclusions are mentioned where relevant, these data will be published under separate cover. Flight morphology is predictive of flight style and flight ability, and, therefore, habitat use of birds and bats (Norberg, 1986; Norberg and Rayner,
1987). Characters such as wing area in relation to body mass, length and width of
10 wings, and shape of the wing tips influence speed and maneuverability. Bats that forage in open areas tend to fly fast. Bats that forage in closed areas require maneuverability rather than speed (Vaughan, 1966; Norberg, 1986; Norberg and
Rayner, 1987). Sympatric North American bats vary in morphology and habitat.
Character release in the Hawai\ ian hoary bat should thus be reflected in more intermediate wing parameters which enable the use of both open and closed habitats.
Similarly, relatively more robust skulls in insectivorous bats have been related to dietary specializations (Freeman, 1979; 1981). Bats with thicker dentaries, higher coronoid processes, reduced maxillary tooth rows, and larger masseter and temporalis muscles, are able to take both hard-bodied (e.g., beetles) and soft-bodied insects (e.g., moths)(Wamer, 1985). Bats with more gracile skulls specialize on soft-bodied insects, like moths (Freeman, 1979). The North
American hoary bat feeds predominantly on moths (Ross, 1967; Black, 1972; 1974;
Barclay, 1985; Warner, 1985). Presumably, the diet and related morphology of the North American hoary bat were the product of natural selection to minimize interspecific competition among North American insectivorous bats. Divergence in the skull morphology of the Hawai \ian hoary bat would thus be expected to be in the direction of a more robust skull for greater utilization of hard-bodied insects.
Any attempt to interpret divergence in the Hawai \ian hoary bat as the result of character release is difficult because it relies upon the demonstration that the evolution of a population of organisms has been permitted by an absence
11 or reduction of interspecific competition in the past. Since we cannot reconstruct
events of the past precisely, character release must be inferred from present
events by evaluating the assumptions upon which it rests and the predictions it
generates (Grant, 1986).
The character release hypothesis consists of a number of assumptions and a
major prediction. The three assumptions of the hypothesis are: (1) morphological
change has occurred, (2) the morphological change was an adaptive result of
natural selection operating on heritable characters, and (3) the morphological
change was permitted by the absence or reduction of interspecific competition.
The major prediction of the hypothesis is that some or all aspects of the foraging
niche of the organism has expanded to include the niches of its ancestral
population and that of the putative competitors of the ancestral population in the
zone of sympatry. Such expansion is only possible if permitted by the existence of
vacant niches. In this instance vacant niches are defined with respect to putative
competitors of the ancestral population in the zone of sympatry.
ORGANIZATION OF THE DISSERTATION
There is little published information on the Hawai>ian hoary bat
pertaining to character release. The studies of Baldwin (1950), Tomich (1974,
1986), Whitaker and Tomich (1983), Belwood and Fullard (1984), Fullard (1984), and Fujioka and Gon III (1988), help to raise the issue of character release, and document sightings of the bat. However, knowledge of the ecological context of
12 the distribution and abundance of this bat is inadequate. Chapter 2 deals with distribution and abundance of the Hawai\ ian bat on the island of Hawai \ i. It details the geographical as well as the temporal aspects of this bat I s distribution.
Information on the location of foraging sites of this bat and the elevational, temperature, and rainfall range over which these are distributed, is also provided.
The abundance of the Hawai \ ian bat at each of these sites was estimated, and the relationship between bat activity patterns and insect biomass was investigated.
Chapter 3 deals with the morphological divergence of the Hawai \ ian bat.
In this chapter, I show that divergence has occurred in the flight and skull morphology of the Hawai\ ian hoary bat. I then evaluate the functional significance of this divergence by considering mechanical implications of the divergence, and by comparing the morphology with that of other species of insectivorous bats.
Chapter 4 demonstrates the significance of the morphological changes in the Hawai \ ian hoary bat through a consideration of the relationship between foraging behavior and diet. This is done in the context of the differential detectability hypothesis. This hypothesis asserts that apparent selection of large prey by fast flying bats, which must use a low frequency echolocation signal to increase their detection range, is the result of a reduction in the detectability of smaller prey. The interdependence of flight and echolocation for increasing the net rate of energy gain of the animal, and for varying handling time in response to changes in prey density, are also investigated.
13 In Chapter 5, I attempt to evaluate the morphological and ecological divergence in the Hawai \ ian bat in terms of character release. I test the prediction that the Hawaiian bat has a more diverse diet than its North American counterpart, and that its diet is a combination of the diets of its North American counterpart and other mainland insectivorous bats which are putative competitors of the North American hoary bat. I also discuss the problems encountered when trying to infer historical phenomena like character release from an analysis of present events.
14 CHAPTER 2
DISTRIBUTION AND ABUNDANCE OF THE ENDANGERED HAWAI'IAN
HOARY BAT, LASlURUS ClNEREUS SEMOTUS, ON THE ISLAND OF
HAWAI'I
INTRODUCTION
The Hawaiian hoary bat, Lasiurus cinereus semotus, is the only extant species of bat in the Hawaiian archipelago. Although it has been on the endangered species list since 1970 (Tomich, 1986), no recovery plan exists. This is partly because very little is known about the natural history or the distribution and abundance of this bat. Information on its distribution is limited to reports of incidental sightings on the island of Hawai \ i (Perkins, 1913; Baldwin, 1950; Bryan,
1955; Tomich, 1986) and Maui (Duvall and Gassman-Duvall, 1991), or comes from surveys peripheral to surveys conducted for forest birds (Kepler and Scott,
1990). The only focused survey on the distribution and abundance of the
Hawaiian hoary bat was the echolocation census conducted by J. H. Fullard
(unpub. manus.) on Kaua\ i. The little that is known about the distribution and abundance of the Hawaiian hoary bat is confusing. For example, it is reported as both common (Perkins, 1913) and rare (Baldwin, 1950; Kepler and Scott, 1990) at higher elevations on the island of Hawai \ i.
15 The purpose of this study Vias to locate areas on the island of Hawai ' i where bats occurred regularly and to estimate the abundance of this bat in such areas. Both visual observation and echolocation censuses were used.
MATERIALS AND METHODS
The survey was undertaken on the island of Hawai 'i over a three year period (1990 to 1992) for a total of 360 hours spread over 101 days. Bat echolocation calls were monitored by driving at speeds of 25-40 kmh" along roads accessible to two-wheel drive vehicles (Figure 2.1). The echolocation calls were monitored using a QMC Mini Bat Detector (QMC Instruments Ltd., 229 Mile
End Road, London El 4AA, England) tuned to a frequency of 30 Khz, which is the peak frequency of the Hawaiian bat I s echolocation calls (Belwood and
Fullard, 1984). Each section of road was traversed at least twice (once in each direction) between 1830h and 2030h each night of survey. Sections of road were surveyed on at least two successive nights during September and October of 1990.
The direction that surveys were begun on successive nights were reversed to reduce the effects of time of night.
Visual observations were also carried out at localities selected on the basis of the type of habitats in which bats were previously seen foraging (Tomich, 1986,
HHP, 1992). Surveys at these sites (Figure 2.1) commenced at dusk and lasted until dark, a period of about one hour. The numbers of bats at each site was conservatively estimated by counting the maximum number of bats that could be
16 seen at anyone time. Counts were made every 10 minutes and the largest count
in the hour of the survey recorded. Since these bats dart in and out of foliage and
one's field of view, they are not all visible at the same time.
Three of the sites (Kipahoehoe, Red Cinder Road, and Ocean View) were also surveyed after dark using a bat detector. During this time surveys were conducted every half hour from 1900h to 2400h, and every hour from 2400h to
0600h, for two nights at the first two sites and six nights at Ocean View. When using the bat detector each detection was counted as a single bat. A fourth site
(Waiono Ranch) was also surveyed from just before dawn to about one hour after sunrise, on two consecutive days.
Sites where bats were seen on the first night were surveyed at least once again. With the exception of Ocean View which was surveyed from May 1991 to
October 1992, Pohakuloa which was surveyed from May to October 1992, and
Volcanoes National Park which was surveyed in October 1992, all sites were surveyed between September and October 1990, and from February to August the following year. The insect fauna at Ocean View was sampled by means of a 22 watt ultra-violet insect light trap (BioQuip Products), from February to October,
1992.
17 RESULTS
VISUAL OBSERVATION
Bats were seen at 16 of the 26 sites visited (Figure 2.1). Two of the 16 sites were visited once only. These were located at Pu 'uhonua 0 Honaunau Historical
Park on the leeward coast (maximum count of two bats in one hour of visual observation) and at the entrance to the Kulani Correctional Facility (maximum count of 3 bats in one hour of visual observation) at the end of Stainback
Highway. At the remaining 14 sites bats were observed on two or more consecutive nights, during September and October (Table 1). They can therefore be regarded as sites used regularly by the bats (at least at some times of the year - see Temporal Patterns).
The largest number of bats seen in the hour from dusk to dark at each of the 14 sites is a conservative estimate (Table 2.1). For example, at the Red Cinder
Road site (at about the 70 mile marker on Highway 11), the maximum number of bats observed was 5, but 12 different individuals were captured at this site over 14 trap nights.
In September 1990 I saw a total of 15 bats flying, singly or in pairs, upslope from below Highway 11 over Kipahoehoe Natural Area Reserve, and disappearing over the horizon high up the slope of Mauna Loa. Similar sightings have been made over the Manuka Natural Area Reserve about ten miles to the south of
Kipahoehoe (Fujioka and Gon III, 1988).
18 ECHOLOCATION MONITORING
Almost all of the bats detected by echolocation monitoring were
concentrated south of Kailua (Kona) between the 80 and 96 mile markers along
Highway 11 (Figure 2.1). A total of 18 bats (i.e. 18 detections) were detected between the 80 and 96 mile markers over two nights of survey (4.25±1.78 bats per
traverse). Not a single observation or echolocation detection was made in N. W. and S. W. Kohala on the leeward side of the island. However, 40% of all bat sightings occurred on the drier leeward side of the island south of Kailua, compared to 26% on the wetter windward side.
CONTEXT OF OBSERVATIONS
Ecological
The bats were all in flight and obviously foraging at the 16 sites where they were seen. They displayed the darting, weaving flight, with sudden turns and dives, characteristic of foraging insectivorous bats that catch their prey on the wing. Bats detected (using the bat detector) along Highway 11 may have been foraging as well. Visual observations of foraging bats prior to nightfall along Highway 11, especially those parts of the road at or near Kipahoehoe and Manuka Natural area Reserves, suggest that the bats detected after nightfall along this section of road may have been foraging as well.
19 The 15 bats observed flying over Kipahoehoe Natural Area Reserve in
September 1990 appeared to be in transit. These bats were flying very high (> 150 m) and did not display the darting and weaving flight characteristic of a foraging bat.
Temporal Patterns
Diurnal activity patterns
The commencement and duration of foraging activity varied slightly from site to site. For example, during September and October (1990 and 1991) bats started foraging at Red Cinder Road at 1800h or immediately thereafter. At the
Ocean View and Kipahoehoe sites, on the other hand, they first appeared between 1830h and 1845h. At the latter two sites the bats remained active until about 0200h. There was, however, a peak in activity just before sunset (four bats visible in the hour from dusk to dark) which dropped slightly as the night progressed (two bats detected per hour). At Red Cinder Road activity came to an almost complete stop at about 190Gh. Out of five nights of surveying only one bat was detected briefly after 1900h - at 2100h on two separate occasions. Pohakuloa
(Bradshaw Military Airfield) was different in that there did not appear to be a drop in activity. The number of bats remained constant at least until 0230h when observation ceased.
Early morning foraging was observed. Between three to five bats were observed on two consecutive mornings foraging just before dawn and for at least
20 45 minutes after sunrise at Waiono Ranch (Kailua - Kona). At Kipahoehoe, the only other site at which bats were observed foraging at dawn, a single bat was seen on one of the two mornings of sampling. Early morning foraging in bats has also been reported on Maui (Duvall and Gassman-Duvall, 1991).
Seasonal activity patterns
There appeared to be a drop in activity at most foraging sites between
February and August (Table 2.2). At Kipahoehoe, Manuka, and Red Cinder Road bat sightings dropped to 3 bats over 3 nights of sampling. Bats did not spend much time at these sites. Most of them made one to three foraging circuits before disappearing. At Ocean View sightings dropped to one bat over 4 nights of sampling in February, and to 4 bats over 3 nights of sampling in March. From
May to August the numbers of bats increased at this site and returned to the
September/October (1991) levels (Table 2.2), with the exception of a decrease in activity during July and October (Figure 2.2). The drop in activity at Ocean View seems to be correlated with a drop in insect biomass (Figure 2.2). In the
Multipurpose Range Complex of the Pohakuloa Military Training Area, along the
Bobcat trail (1 800 m elevation), between one to three bats were seen and/or detected over a sampling period of four nights for each month between February and August 1992 (Table 2.2). This number increased to 11 bats over four sampling nights in August (HHP, 1992). Incidental sightings of bats at Hakalau Forest
National Wildlife Refuge (1 890 m elevation) by Jaan Lepson (pers. comm.)
21 follow the same pattern. Although bats are sighted between January and August, most sightings occurred from September to December. No bats were seen or detected at any of the other sites between February and August during the present survey (Table 2.2). Bat numbers remained constant at Pohakuloa
(Bradshaw Military Airfield) from May to October.
ENVIRONMENTAL CORRELATES
Habitat characteristics
The 14 foraging sites can be divided into four habitat categories: native vegetation, exotic vegetation, mixed (native and exotic) vegetation, and open ocean (Table 2.3). Of the 81 bats observed at these sites, 44% were associated with native vegetation, 16% with exotic vegetation, 9% with mixed vegetation, and
25% foraged over open ocean (Table 2.1). Bats were therefore associated more with native vegetation than with exotic vegetation, contrary to what Kepler and
Scott (1990) found.
Temperature
The foraging sites of the Hawai' ian bat extend over an elevational range of 0 - 1 830 m (Table 2.3). The climatic conditions experienced by these bats therefore range from tropical to temperate. At elevations below 1 500 m temperatures above 35°C are infrequent and the range between the coldest and
22 warmest months averages 6.5°C. The average temperature at Hilo (sea level) in
August is 24.1°C and in January 21.6°C. The average temperature at Mauna Loa
Observatory (3 398 m) in January is 5°C and in August is 8°C (Blumenstock and
Price, 1972). However, at times the upper slopes and summits of Mauna Loa and
Mauna Kea are covered in snow during winter months. Cold air formed
immediately above this snow cover descends into areas like Pohakuloa and
Hakalau causing temperatures as low as -15°C (Blumenstock and Price, 1972).
The Pohakuloa and Hakalau sites are at an elevations of about 1 830 m and 1 890
m, respectively. The result is that these sites often experience frost during winter.
Rainfall
Bats were observed most frequently in areas with moderate rainfall. Eight
of the 14 foraging sites had a mean annual rainfall ranging from 1 016 mm to 1
905 mID. Four had a mean annual rainfall ranging from 1 905 to 2 540 mm.
Pohakuloa had a mean annual rainfall of 508 mm.
Bats at Red Cinder Road and Kipahoehoe sites, generally were not active when it rained. In fact on nights when it rained just before or just after the time the bats usually appeared at these sites, the foraging aggregations did not appear and at most one or two bats were seen. However, at Kalopa State Park on the windward side of the island, two bats were observed foraging in a light rain for at least 30 minutes.
23 Wind
In general there was little to no wind at the inland foraging sites. The tall vegetation characteristic of most of these sites provided a natural wind break.
However, these bats are fairly strong flyers. I have observed them flying without much difficulty in fairly strong winds. At Pohakuloa, where the vegetation was predominantly low mamani-naio (Sophora chrysophylla and Myoporum sandwicense) forest, wind speed ranged from 3 - 6 kmh', occasionally gusting to
11 kmh", on 60% of the nights surveyed.
DISCUSSION
Visual observations and echolocation monitoring suggest that the Hawi\ ian hoary bat is fairly widespread around the island of Hawai \ i. Although bats appear to be more abundant on the drier leeward side of the island, they may not be as rare on the windward side of the island as previously thought (Kepler and Scott,
1990). The largest foraging aggregation (11 bats) was seen in September 1990 over Keokea Bay on the windward coast. Large groups of bats have been reported on the windward side of the island - 12 bats over Hilo Bay in September 1963 and
22 bats at Honoka\ a landing in October 1964 (Tomich, 1986). A more extensive survey of this coastline needs to be undertaken. I have received reports from surfers of bats foraging over water all along this coast.
There appears to be a decrease in bat activity from February to August
(Table 2.2). The drop in activity occurs over a wide elevational and geographical
24 range. In addition, bats are present as early as May at Ocean View and Pohakuloa
(Bradshaw Military Airfield), two sites with very different elevations and geographical locations. These patterns argue against an elevational or geographic migration within the island of Hawai\i.
The drop in bat activity with insect biomass at Ocean View suggests that these bats may use a number of foraging sites over the cause of a night or a few days. The particular sites used depend on local insect biomass. Radio telemetry data (Jacobs unpub.) suggest that a bat may stay away from a previously visited site for up to two weeks. If this is the case, the apparent drop in activity from
February to August at some sites may be associated with times when insect biomass was low at these sites and bats were foraging elsewhere.
The apparent increase in bat activity in September/October (in terms of the number of sites at which bats are seen) may be the result of young of the year fledging. Capture data (Jacobs unpub.) suggest that lactation occurs in June and
July and young are fledged sometime in August.
In conclusion, the numbers of bats reported here for the different sites at different times of the year suggest that the Hawai \ ian hoary bat is rare and that its endangered status is appropriate. The fact that the habitats used by this bat are more often than not associated with native vegetation suggests that the major threat to its existence in Hawai \ i may be the destruction of native vegetation.
Protection of native vegetation is, therefore, essential to the survival of Lasiurus
25 cinereus semotus in Hawai ' i. This may become even more evident when the
roosting ecology of this bat is studied.
26 CHAPTER 3
MORPHOLOGICAL DIVERGENCE IN THE HAWAI'IAN HOARY BAT
LASIURUS ClNEREUS SEMOTUS
INTRODUCTION
Bats are the only mammals to have reached isolated oceanic islands (e.g.
Hawaivi and the Galapagos) without human assistance (Darwin, 1950; Carlquist,
1965). The Hawai ' ian islands, situated 4 000 km from the nearest continent and 1
600 km from the nearest archipelago (Simon, 1987), have one extant and one fossil species of bat (Tomich, 1986; Ziegler and Howarth, unpub. mans.). The extant Hawai' ian hoary bat, Lasiurus cinereus semotus, is thought to be derived from the North American hoary bat, Lasiurus cinereus cinereus (Sanborn and
Crespo, 1957; Tomich, 1986; Fullard, 1987). Previous studies suggest that the
Hawai'ian bat has diverged morphologically from its ancestral population: it has undergone a reduction in body size (Tomich, 1986; pp.23), and appears to have a more generalized diet than that of its North American counterpart (Whitaker and
Tomich, 1983). This study provides a more comprehensive analysis of the morphological divergence.
Body size is a fundamental morphological trait that is correlated with many other morphological, ecological, physiological, and life history traits (Peters, 1983;
Schmidt-Nielsen, 1984; LaBarbera, 1989; Riess, 1989). It has implications for characteristics as basic as size of prey consumed, or the kind of habitat used.
27 Larger animals are able to take both small and large sized food while smaller animals cannot switch to larger food items if their preferred prey declines in abundance. Instead, they have to respond by hibernation or migration to areas where the appropriate prey can be found (Peters, 1983, pp.114). In addition, body size influences both the size of foraging area and the type of habitat used.
Maximum speed of locomotion increases with size, and this allows an increase in foraging area (Peters, 1983, pp.86-90). Within each locomotive mode (running, swimming, flying), larger animals are better commuters than smaller ones and therefore are able to utilize greater foraging areas (Pennycuick, 1979; as cited in
Peters, 1983). Small body size, on the other hand, may permit existence where resource supply is low. For example, freshwater midges decrease in mean adult size in less productive waters (Davies, 1980), and islands that have fewer total resources generally have smaller-sized animals (Brown, 1971).
Divergence in body size has implications for the size and shape of morphological structures. The relationship between body size and a morphological trait can be one of geometric similarity in which there is a change in body size without a change in shape. Geometric similarity usually results in a change in function over a wide range of body sizes because weight increases at a greater rate than does cross sectional area (Gould, 1975; LaBarberra, 1989). Under geometric similarity, the slope of the line plotting two linear dimensions on double logarithmic coordinates would be 1 ([I X [1). Similarly, since area varies as the square of the linear dimension and volume as the cube of the linear
28 dimension, a plot of a linear dimension and volume (e.g. mass, since mass is
proportional to volume) would have a slope of 0.33 ([1 x P), and a plot of area
against volume will have a slope of 0.67 (z2 x P).
Changes in body size either during growth and development, or within or between taxa, usually result in changes in shape (Gould, 1971; 1975; Fairbairn,
1992). Such departure from geometric similarity is called allometry. Allometry between body size and another morphological trait can be expressed in the form of the power function
y=~ where y is the morphological trait in question, M is body mass (or some other measure of size), and ex and B are fitted constants. When regressed on log-log coordinates the function is a straight line with slope Band y-intercept ex.
Allometry usually occurs so that functional equivalence over a wide range of sizes is maintained (e.g. allometry in mammalian teeth - Gould, 1975). It is possible, however, that function may be sensitive to changes in shape in which case an allometric change in shape results in a change in function. Such a case can be found inmorphological traits which have aerodynamic or hydrodynamic properties that vary with body size. Under geometric similarity there is no change in shape and these parameters are not affected.
Body size directly affects aerodynamics of birds and bats through wing loading, a measure of wing area relative to body mass, and induced power, the work required to generate lift. Wing loading is positively correlated with flight
29 speed and negatively correlated with maneuverability (the ability to tum tightly) and agility (the ability to execute turns at high flight speed). This direct effect is the basis for change in function of other wing parameters that may be indirectly affected by body size. These include aspect ratio (the ratio of wingspan to wing area) and tip shape index (a measure of how rounded or pointed the wing).
Aspect ratio is positively correlated with flight cost efficiency and flight speed
(Findley, et al., 1972; Norberg, 1981; Norberg and Rayner, 1987; Norberg, 1990).
Tip shape index is positively correlated with lift.
Wing loading, aspect ratio, and tip shape index determine flight style, flight ability, and therefore habitat use of birds and bats (Norberg, 1986; Norberg and
Rayner, 1987). Species with high wing loading, high aspect ratio, and relatively pointed wings fly faster and are less maneuverable than species with low wing loading, low aspect ratio, and more rounded wings (Figure 3.1). The former species would be specialized for flying in open habitats (away from vegetation) where space is not limited and a high degree of maneuverability is not advantageous. The latter species would be specialized for foraging in closed habitat (within or near vegetation) where the ability to maintain lift at low speeds and increased maneuverability are advantageous (Vaughan, 1966; Norberg, 1986;
Norberg and Rayner, 1987). Induced power, the main power in slow flight, is directly proportional to body size and inversely proportional to wingspan squared
(Norberg, 1990, pp.25). Thus, a bat capable of slow flight, necessary for the
30 exploitation of closed habitats, would be expected to have a relatively small body size (mass) and a relatively long wingspan.
Changes in shape have also brought about changes in function in the skull morphology of bats. Relatively more robust skulls in insectivorous bats have been related to dietary differences (Freeman, 1979; 1981). Bats with thicker dentaries, higher coronoid processes, reduced maxillary tooth rows, and larger masseter and temporalis muscles are able to eat hard-bodied insects, like beetles, because of the increased crushing power that these features confer upon the jaws of these bats (Maynard Smith and Savage, 1959). For example, the temporalis muscle, which acts to close the jaw, originates on the cranium and inserts on the coronoid process of the dentary. The condyle of the dentary articulates with the rest of the skull and represents the fulcrum of the jaw. An increase in the height of the coronoid process increases the moment arm of the temporalis muscle about the condyle. This in turn increases the crushing power of the jaws (Maynard Smith and Savage, 1959). Since bats with more robust skulls are able to eat hard-bodied insects, they are also able to take soft-bodied insects (Warner, 1985). On the other hand, bats with more gracile skulls specialize on soft-bodied insects (Freeman,
1979; 1981).
Here I investigate divergent evolution through geometric and allometric analysis of flight and skull morphology of the Hawai \ ian and the North American hoary bats. An ecological and a comparative approach are used to interpret changes in function.
31 METHODS
SKULL MORPHOLOGY
Skull measurements were taken from museum skull specimens (courtesy of the Museum of Southwestern Biology, New Mexico) of the two hoary bat subspecies (Figure 3.2; Appendix, Table A.l). Measurements for each specimen were taken with a dial calipers under a Zeiss dissecting microscope (15x magnification). The occlusal surface of the maxillary tooth row (from the fourth premolar to the third molar) for each specimen was drawn under a Wild M5 camera lucida. The area of the maxillary tooth row was measured from this enlarged drawing using a Jandel digitizer with SigmaScan software (version 3.10).
Measurement error was estimated by taking repeated measures of some measurements from a single skull. In all cases measurement error was much less than the difference between the means of the two subspecies (Appendix, Table
A.2). Masses for each specimen were obtained from museum records.
FLIGHT MORPHOLOGY
Wing design was described in terms of the following parameters (Figure
3.3):
Wing span, b, is the distance between the wing tips of a bat with
extended wings.
Wing area, S, is the area of the two wings, the tail membrane, and the
body, excluding the head.
32 Wing loading, B = Mg/S,
where M = total body mass
g = gravitational acceleration
Aspect ratio, A = b2IS,
Tip shape index, I = Ts/(T( - Ts), where T, = tip area ratio = Shw/Saw
T( = tip length ratio = lhw/law
Morphological measurements of the Hawai>ian hoary bat were taken from
35 live adult specimens (24 females and 11 males) caught in mist nets operated at
Ocean View, Red Cinder Road, and Pohakuloa on the island of Hawai\ i
(Chapter 2). Mass (to the nearest 0.5 g) and forearm length (to the nearest 0.1 mm) were measured for each individual caught. Forearm length and mass were also obtained from museum specimens, courtesy of the Bernice P. Bishop
Museum (Honolulu) and the Museum of Southwestern Biology (New Mexico).
The right wing of each captured bat was traced so that the following measurements could be taken using a Jandel digitizer with SigmaScan software
(version 3.10): armwing length (law), handwing length (lhw)' wing span, armwing area (Saw)' handwing area (Shw), and total wing area. These wing measurements were used to calculate wing loading, aspect ratio, and tip shape index. Wing parameters for 11 female and four male L. c. cinereus were obtained from
Norberg and Rayner (1987, one female), and from wing tracings provided by
Catherine Koehler (10 females and 3 males), and Matina Kalcounis (one male).
33 Aspect ratio and tip shape index were compared directly through t - tests,
since they describe the shape of wings and wing tips independently of mass
(Norberg and Fenton, 1988).
STATISTICAL ANALYSIS
Analysis of covariance was performed on skull measurements, wing span,
wing area, and wing loading, with skull length and body mass as covariates for the
skull and wing measurements, respectively, and with species and sex as factors.
This tested the possibility that differences in the degree of sexual dimorphism
existed in the two populations (Kleinbaum, Kappa, and Muller, 1988). A least squares regression of each of wing span, wing area, and wing loading on body
mass was then performed on each population separately (because their masses did
not overlap), and on the combined data for the two populations (for goodness of fit of a single line). Similar regressions were also performed for each of the skull measurements on skull length for the combined populations (skull lengths of the
two populations overlapped slightly). Slopes thus obtained were compared with
that predicted by geometric similarity through the 95% confidence interval of the estimated slope. Only female bats were used because there were too few males for comparison. Sample sizes were 24 and 12 for the flight and skull measurements, respectively, for the Hawai \ ian hoary bat, and 11 for both the flight and skull measurements for the North American hoary bat. All data used in
34 regression analyses were log-transformed (base 10) and analyzed using SAS/STAT
(version 6) statistical software on a microcomputer.
FORAGING BEHAVIOR
Observation of the foraging behavior of the Hawaiian hoary bat was conducted at eight sites on the island of Hawai \ i: Keokea Bay, Waimanu Bay,
Waiono Ranch, Pohakuloa, Ocean View, Red Cinder Road, Kipahoehoe and
Manuka Natural Area Reserves. Foraging behavior was observed for approximately one hour at dusk each day while sufficient light remained.
Observations were concentrated at Ocean View, Red Cinder Road (44 hours over
11 months each), and Pohakuloa (18 hours over six months). Foraging behavior at each of the other sites was observed for two hours spread over two days. Flight speeds of the bats were measured at Pohakuloa and Ocean View using stop watches to record the time it took foraging bats to fly straight paths between two predetermined points a known distance apart.
RESULTS
Since there were differences between sexes within each subspecies (females are larger than males in both subspecies - Table 3.1), sexes were analyzed separately. Male and female Hawai \ ian bats were smaller than their North
American counterparts with respect to mass (t = 16.31, DF = 16, P < 0.001; t =
16.7, DF = 32, P < 0.001, respectively) and forearm length (t = 6.98, DF = 19, P
35 < 0.001; t = 13.2, DF = 37, P < 0.001, respectively - Table 3.1). Female
Hawai \ ian hoary bats were approximately 45% smaller than their North
American counterparts.
Analysis of covariance yielded no species by sex interaction for body mass
(F = 2.06, DF = 48, P = 0.16), wingspan (F = 0.42, DF = 48, P = 0.52), wing area (F = 0.02, DF = 48, P = 0.89), wing loading (F = 1.87, DF = 0.48, P =
0.18), or for any of the skull measurements (F < 3.8; DF = 39, P > 0.05). Thus, there was no difference in the degree of sexual dimorphism in the two subspecies.
Further analyses were restricted to data for female bats.
SKULL MORPHOLOGY
Divergence is evident in the skull morphology of the Hawai \ ian bat.
Absolute mean skull measurements for the two subspecies show that the smaller
Hawai \ ian bat has measurements nearly as large, and in some cases larger (e.g., the height of the coronoid process), than that of its North American counterpart
(Table 3.2).
The Hawai \ ian bat has a relatively longer skull. The slope for body mass regressed against skull length was lower than that for geometric similarity (Table
3.3, Figure 3.4). The slopes for other skull measurements regressed against skull length are not distinguishable from those for geometric similarity (Table 3.3), since the 95% confidence limits for these slopes encompass the respective slopes for geometric similarity. This suggests that, although there has been a relative
36 increase in the overall size of the Hawai \ ian bat's skull, the proportions of the skull have remained the same as that of its North American counterpart.
However, there are important exceptions. These include the a:b ratio (the ratio of the distance from the jaw joint to the origin of the masseter muscle, to the distance from the joint to the insertion of the masseter muscle), the length of the masseter muscle scar, and the height of the coronoid process. These are all larger in the smaller Hawai \ ian hoary bat (Table 3.2). This is also true for the height of the coronoid process above the condyle (0.7±0.1 mm and 0.3±0.1 mm for the Hawai \ ian bat and the North American hoary bat, respectively; t = 3.03,
DF = 16, P < 0.008), obtained by subtracting the height of the condyle from the height of the coronoid process. These measurements were taken from the same reference plane (Figure 3.2a). As will be discussed below, these differences enable the Hawai \ ian hoary bat to consume relatively larger and more hard-bodied insects.
FLIGHT MORPHOLOGY
Divergence is also evident in the flight morphology of the Hawai \ ian bat.
It has a smaller absolute wingspan and wing area, and lower wing loading than its
North American counterpart (Table 3.4).
Least squares regressions yielded significant slopes for wingspan (North
American hoary bat only) and wing loading (both subspecies) when the subspecies were considered separately (Table 3.5). The slopes and intercepts for wing loading
37 for the two populations were similar suggesting that a single regression could be fitted to the combined data for the two populations. Because significant slopes with higher correlation coefficients could be obtained for all wing parameters when the two populations were considered together, further analyses involve the combined populations.
The slopes were lower from what would be expected from geometric similarity (Table 3.5). This indicates that the Hawai \ ian bat has a relatively longer wingspan and greater wing area than its North American counterpart
(Table 3.5, Figures 3.5 and 3.6). This means that it has a relatively lower wing loading as indicated by a slope which is steeper than that for geometric similarity.
(Table 3.5, Figure 3.7). The decrease in body mass per se also means that this bat has lower induced power at lower flight speeds. As will be discussed below, these changes enable the Hawai\ ian bat to fly slower and with greater maneuverability.
Aspect ratios and tip shape indices of male and female Hawai \ ian hoary bats are not statistically different from those of their North American counterparts (Table
3.6).
The fact that a single allometric line could be fitted to both populations indicates that the divergence in the flight morphology of the Hawai \ ian bat is probably the result of a direct change in body size with correlated allometric responses in the size of the wings (Table 3.5, Figures 3.5, 3.6, and 3.7).
38 FORAGING HABITAT AND HEHAVIOR
The North American hoary bat is an aerial hawker of flying insects. It forages almost exclusively in open habitats, away from vegetation and other obstacles (Barclay, 1985). In these open habitats it uses rapid, relatively unmaneuverable flight with low frequency echolocation calls designed to detect insect prey at a distance (Barclay, 1985; 1986). In contrast, the Hawai'ian bat utilized both relatively closed habitats near vegetation and open habitats including space above the tree canopy and over open ocean.
The open habitats included Keokea Bay, Waimanu Bay, Waiono Ranch, and Pohakuloa (Chapter 2). At Keakeo Bay and Waimanu Bay, between 5 to 15 bats were observed each night. Here bats foraged over the water in excess of 100 m offshore, at heights ranging from about ten to a 150 meters. At Waiono Ranch, vegetation consisted of native 0'hia (Metrosideros polymorpha) forests interspersed with large tracts of grass lees. Up to 7 bats commonly foraged at least 50 meters above the tree canopy and were never closer than 20 meters to any vegetation. At
Pohakuloa, between seven to ten bats were observed each night foraging in an area illuminated by a mercury vapor lamp attached to the top of a 9 m pole. This light illuminated the petroleum, oil, and lubrication depot of Bradshaw Military
Airfield. Vegetation in the area consists mainly of open mamane-naio (Sophora chrysophylla and Myoporum sandwicense) woodland (Sohmer and Gustafson, 1987).
There was no vegetation higher than 6 m in the area. The bats commonly foraged at heights of 15 m, or more.
39 In these habitats the Hawai' ian bat used the same foraging behavior as
that of its North American counterpart. Its flight was rapid and relatively unmaneuverable and it used a low frequency echolocation signal, albeit not as low as that of its North American counterpart (James Fullard, David Jacobs, and
Robert Barclay, mans. in prep.), The mean flight speed of the Hawai'ian bat in
the open habitat at Pohakuloa (open habitat) was 11.39±2.48 ms", similar to the flight speed of 11.00 ms" recorded for this bat in an open habitat (bats foraged
around lights removed from vegetation) on the island of Kaua vi (Belwood and
Fullard, 1984). Quantitative data on the flight speeds of free ranging, foraging
North American hoary bats are not available.
The closed habitats included Ocean View, Red Cinder Road, and
Kipahoehoe Natural Area Reserve (Chapter 2). In these habitats the foraging behavior was characterized by slower more maneuverable flight, and a higher frequency echolocation signal (Fullard, Jacobs, and Barclay, mans. in prep.). The mean (±SD) flight speed at Ocean View was 6.56±0.90 ms", just over half that in the open habitat. These sites consisted of clearings in lowland mesic 0'hia
(Metrosideros polymorpha) forest. Between five to ten bats would often fly up and down in these clearings in oblong circuits, occasionally darting in and out of the
0'hia foliage at the sides. They commonly flew at heights ranging from about 1 to
15 meters above ground, well below the forest canopy which extended to about 25 meters. The width of the clearings ranged from about 10 to 20 meters. When flying down the center of these clearings bats were between 5 to 10 meters from
40 the nearest vegetation, but regularly came within one meter of the foliage when flying through the vegetation or pursuing insects. This is the basis for designating these sites as closed habitats. Sudden turns and dives were more common than in the open habitats. Flight in general was more maneuverable and butterfly-like, in contrast to the swallow-like flight observed in the open habitats.
DISCUSSION
Divergence in the Hawai \ ian hoary bat included characters related to flight and feeding. The reduction in body size was associated with allometric responses in wing size and skull length. The decrease in body size per se had marked aerodynamic implications. It permitted a decrease in wing loading and induced power without altering the high aspect ratio of the ancestral subspecies.
The relative increase in the size of the skull was accompanied by non-allometric changes in the a:b ratio, size of the masseter muscle, and height of the coronoid process. Here I discuss the functional significance of this divergence through evaluation of the mechanical implications of the changes and through comparisons with other species of insectivorous bats.
The jaws of insectivorous bats have characteristics of both carnivores and herbivores. They need the crushing power of carnivores for subduing live insect prey and piercing the hard insect exoskeleton. They also require some lateral movement of the jaw for mastication of the insect exoskeleton (Freeman, 1979). A larger gape is required for larger insects.
41 The temporalis and masseter muscles affect power and movement of the jaws. The temporalis muscle, which acts to close the jaw, originates on the cranium and inserts on the coronoid process. The condyle of the jaw articulates with the s1.."1111 and represents the fulcrum of the jaw. An increase in the height of the coronoid process increases the moment arm of the temporalis muscle around the condyle (Maynard Smith and Savage, 1959). The masseter muscle originates from the area in front of the zygomatic arch and inserts on the dentary as can be seen from the scar on the dentary (Figure 3.2c). The main function of the masseter is the lateral masticatory movements of the jaw (Maynard Smith and
Savage, 1959). The a:b ratio is correlated with the gape of the jaws (Freeman,
1979).
The changes in skull morphology thus have functional significance. The relatively larger gape of the Hawai \ian bat, as indicated by the larger a:b ratio, means that the smaller Hawai\ ian hoary bat has retained the ability of its North
American counterpart to eat large insects. The large masseter muscle, as indicated by the longer masseter muscle scar, and the larger moment arm of the temporalis muscle, as indicated by the higher coronoid process, increase the masticatory and crushing power of its jaws. It appears then that the Hawai'ian bat is capable of taking relatively larger, and more hard-bodied insects than its North American counterpart.
The changes in the skull morphology of the Hawai \ ian bat is consistent with the generalization of its diet. Whereas the North American hoary bat feeds
42 predominantly on moths (Ross, 1967; Black, 1972; 1974; Barclay, 1985; Warner,
1985), the Hawai'ian bat has expanded its diet to include hard-bodied insects
(e.g., beetles) while retaining the ability to take large insects despite a marked reduction in body size (Chapter 5, Figures 5.1 and 5.2; Whitaker and Tomich,
1983). Similar relationships between skull morphology and diet of other insectivorous bats have also been demonstrated (Freeman, 1979; 1981). For example, molossid bats that feed mostly on beetles have relatively higher coronoid processes and larger masseter muscles than molossid bats that feed mostly on moths (Freeman, 1979).
The divergence in flight morphology enables the Hawai ' ian hoary bat to add some new patterns to its flight. The lower wing loading of the Hawai> ian bat, combined with the lower induced power at low speeds for this bat, means that it is capable of flying slower and with more maneuverability than its North
American counterpart. Its retention of the high aspect ratio and tip shape index of its North American counterpart means that the Hawai' ian bat has retained the ability for efficient and rapid flight. This increased flexibility in flight behavior has allowed the Hawai ' ian hoary bat to expand its foraging niche to include closed habitats while retaining the ability to utilize open habitats.
The flight of the Hawai'ian hoary bat while foraging in closed habitats is similar to that of other North American insectivorous bats that forage in closed habitats. Lasionycteris noctivagans, for example, forages in clearings among vegetation, and is a highly maneuverable bat capable of efficiently preying on
43 swarms of insects (Barclay, 1985). It has been described as the slowest flying
North American bat with the possible exception of the western pipistrelle
(Barbour and Davis, 1969). Another bat which forages among vegetation is
Eptesicus fuscus pallidus (Barbour and Davis, 1969; Black, 1974). This bat is also a
slow, maneuverable flyer. The foraging flight of this bat has been described as "a steady course broken by frequent sallies to capture insects" (Barbour and Davis,
1969). This is similar to the foraging flight of the Hawai\ ian bat in the closed
habitats.
The association between flight morphology and foraging ecology of the
Hawai \ ian hoary bat is also consistent with that of other insectivorous bats. Its ability to forage in both kinds of habitats, in ways similar to bats that occupy one
or the other kind of habitat in North America, is reflected by its intermediate morphology. The overall mean of the mean male and female wing loadings for the
Hawai \ ian hoary bat and the North American hoary bat are plotted in Figure 3.8.
Also plotted are the overall mean of the mean male and female wing loadings of
E. fuscus and the mean male wing loading of L. noctivagans (Farney and Fleharty,
1969; Norberg and Rayner, 1987). There is no sexual dimorphism in the latter species (Williams and Findley, 1979). The 'morphospace' occupied by the
Hawai \ ian bat is intermediate between that of its North American ancestor (a bat specialized for fast flight in open habitats) and that of L. noctivagans (a slow flying, highly maneuverable bat). The morphology of the Hawai \ ian bat coincides with that of E. fuscus, which has flight patterns similar to it in closed habitats.
44 There is also evidence of convergent evolution between bats and birds with
respect to flight morphology and habitat use (Norberg, 1986).
The comparative evidence supports the functional interpretation of the
morphological divergence of the Hawai \ ian hoary bat. Special attention must be devoted, however, to the marked reduction in body size, which is the largest difference between the two subspecies. A number of explanations have been advanced for insular dwarfism that view resource limitation as the selective agent for smaller body size. Marshall and Corruccini (1978), for example, suggest that insular dwarfism is the result of selective elimination of large individuals, caused by resource limitation, while the within-species gradient (i.e. within-species allometry) is maintained in the rest of the population. Similarly, Lomolino (1985) maintains that the importance of resource limitation in promoting dwarfism increases with an increase in body size of the founder individuals. Limited resources would select for smaller individuals since they require less resources.
Sondaar (1977) suggests that smaller size is advantageous in island environments that are limited in space and resources and in which predation pressure is lower.
Smaller size allows greater mobility, decreases the quantity of food and territorial area required, and increases heat exchange with the surrounding air or water.
According to Bergmann's rule (Scholander, 1955; Smith, 1980), the larger ratio of body surface area to volume of smaller animals facilitates thermoregulation in cases where island populations experience a warmer climate than the parent population. However, this rule is unlikely to account for the small
45 body size of the Hawai \ian hoary bat. It is true that the Hawai\ian bat experiences higher mean annual temperatures than the temperate North
American hoary bat because the major Hawai \ ian islands are situated in the tropics (Blumenstock and Price, 1972). However, the high elevation (4 000 m) to which the island of Hawai \i rises means that some areas experience temperate climates. The habitat of the Hawai\ ian hoary bat covers an elevational range sufficiently broad for it to experience low temperatures similar to those experienced by the North American hoary bat (Chapter 2). Furthermore, the need for more efficient thermal regulation in warmer climates as an explanation for
Bergmann's rule has been called into question (Scholander, 1955; McNab, 1971), and evidence for size-climate relationships in bats does not appear convincing
(Findley and Wilson, 1982).
Selection may have favored smaller body size in the Hawai\ian hoary bat in response to limited resources (Sondaar, 1977; Marshall and Corruccini, 1978;
Lomolino, 1985), but there are complications to this explanation. First, the response to putative resource limitation includes the ability to utilize a wider range of the available resources. It is possible that selection may concurrently have been acting on flight performance. Second, it is not clear that limited resources are a general rule for insular environments. Summed population densities on islands may be less than, comparable to, or higher than continental levels, depending on the particular island, habitat, and organism being studied
46 (MacArthur, et, aI., 1972). The issue is further complicated by the fact that unknown historical, rather than present resources, are pertinent to a selection hypothesis.
In conclusion, the functional significance of divergence in the Hawai \ ian bat is supported by comparative studies with other bats. The correlation between its morphology and ecology appears to be consistent with that in other species of insectivorous bats. This suggests that the divergence is adaptive and potentially the product of natural selection (Darwin, 1950; Futuyma, 1986, pp. 251-253;
Harvey and Pagel, 1991). However, there are problems with the origin of the differences. The changes may have been the result of stochastic processes such as genetic drift and founder effect. They may have been the result of a single founder event or a number of founder events as each island in the archipelago was colonized in tum. The changes in the skull could also have been the result of exposure to hard-bodied insects over the developmental history of the individual bats (Johnston and Gottlieb, 1990). The apparent adaptiveness of these traits could be an incidental result of these processes (Gould and Lewontin, 1979).
Even if differences in wing morphology could be interpreted as adaptations, they might be better viewed as exaptations since they may not have been designed by natural selection for their current function (Gould and Vrba,
1982). Because the divergence took the form of a reduction in body size, with only an allometric response in wing size, one cannot rule out the possibility that
47 changes in the flight abilities of the Hawai \ ian bat, and the associated changes in its foraging niche, were the incidental result of selection acting to reduce body size.
48 CHAPTER 4
FORAGING STRATEGIES AND PREY SELECTION IN THE HAWAI'IAN
HOARY BAT, LASIURUS ClNEREUS SEMOTUS.
INTRODUCTION
Animals have to do more than eat in order to survive and reproduce successfully. For example, they may have to build nests or other kinds of habitable structures, defend territory, provide parental care, migrate, etc. The time and energy spent foraging represent costs in that they detract from the time and energy the animal can spend in other activities. Thus, according to foraging theory, an organism has to make a number of decisions while foraging (where to forage, how to forage, what prey to choose) and these decisions are expected to be shaped by natural selection to maximize the net rate of energy gain while foraging (Charnov, 1976; Davies, 1977; Krebs and Davies, 1978; Pyke, 1981;
Stephens and Krebs, 1986).
An animal making foraging decisions does not have an infinite number of alternatives from which to choose. The choices available to it are limited, or constrained, by the physical environment, other organisms sharing that environment, and its own size, sensory physiology, and locomotion. For example, a bat or bird may be limited in how slow or fast it can fly. Furthermore, choices made may entail some kind of trade-off. For example, if an organism chooses to forage in one mode of locomotion, that choice might restrict the food items available or detectable.
49 Aerial insectivorous bats are appropriate systems for investigating constraints and trade-offs imposed by locomotion and sensory perception. Such bats are hawkers, i.e., they catch flying insects and fly continuously while foraging
(Norberg and Rayner, 1987). They also use echolocation to detect prey at night
(Griffin, 1958). Echolocation is the emission of high frequency sound which is reflected back to ears of the bat as echoes from potential prey. These echoes enable the bat to locate and capture prey.
The echolocation calls of bats can be divided into three phases: a search approach phase, a tracking phase, and a terminal phase (Barclay, 1986). When a bat first detects an insect, it moves from the search-approach phase of its call to the tracking phase. If successful at tracking the insect, it moves to the terminal phase just before making a capture. The shift from one phase to another involves changes in the echolocation call structure which include an increase in pulse rate
(Barclay, 1986).
The combination of aerial hawking and echolocation apparently imposes certain trade-offs upon these bats. Increased flight speed reduces maneuverability
(Norberg, 1986; Norberg and Rayner, 1987), so that faster flying bats must detect prey at a greater range than slower flying bats to allow sufficient time for maneuvering to make a capture. A limitation of echolocation is its short range of detection, typically over a few meters (Kick, 1982). Some bats are able to increase their range of detection by using lower frequency sounds in the search-approach phase of their echolocation. Lower frequency sounds are affected less by
50 atmospheric attenuation than are higher frequency sounds (Griffin, 1971).
However, a drop in frequency results in a decrease in the detectability of smaller insects. Echo intensity declines with a decrease in target size, and the decline is more rapid for targets with diameters smaller than the wavelength of the sound
(Mohl, 1988).
These considerations suggest that fast flying bats using low frequency sound
(long wavelength) either cannot detect small insects, or detect them too late to allow maneuvering for capture. It has been hypothesized that the apparent selection of large insect prey by fast flying bats is a result of the decrease in detectability of small insects, rather than active selection of large insects with higher calorific value (Barclay, 1985; 1986; 1988; Barclay and Brigham, 1991).
This hypothesis, the differential detectability hypothesis, asserts that the size of dietary items of aerial insectivorous bats is determined by the constraints imposed upon them by their particular flying and echolocation abilities.
The differential detectability hypothesis can be tested in an aerial insectivorous bat like the Hawai \ian hoary bat, Lasiurus cinereus semotus
(Whitaker and Tomich, 1983), which uses fast (11.39 ms"), less maneuverable flight in open habitats (away from vegetation or other obstacles), and slow (6.56 ms"), more maneuverable flight in closed habitats (within or near vegetation)
(Chapter 3). In addition, L. c. semotus uses a lower frequency search-approach signal when flying rapidly in open habitats (James Fullard, David S. Jacobs,
Robert Barclay unpublished data). The differential detectability hypothesis makes
51 three testable predictions for this bat. First, it predicts that the diet of the bat will be influenced more by size than by taxon of insect prey. Second, insects taken by the bat in open habitats, where it uses fast flight and a low frequency echolocation call, should be larger than insects taken in closed habitats, where it uses slow flight and a higher frequency echolocation call. Third, it predicts a lower response to smaller than to larger artificial prey when the bat uses fast flight in the open habitat.
If the differential detectability hypothesis is true, it raises the question why
L. c. semotus, which has the option of slow flight with a higher frequency echolocation signal, uses fast flight when foraging in open areas. Fast flight with low frequency search-approach echolocation signals may increase the net rate of energy gain when a bat is faced with an insect fauna in which large individuals are common. Slow flight with a high frequency search-approach echolocation call may increase the net rate of energy gain when larger insects are rare. The differential detectability hypothesis thus has the additional prediction that the insect fauna in the open habitats should consist of a greater proportion of larger insects.
Alternatively, L. c. semotus may be constrained to use slower flight for increased maneuverability to avoid obstacles in the closed habitat, or to use faster flight in the thinner air of the open habitat (higher altititude), independent of differences in insect size.
The nature of prey capture by insectivorous bats makes possible foraging tactics that vary with densities of prey. The bat may be able to decrease its
52 handling time which is the time required to pursue, capture, and consume prey
(Stephens and Krebs, 1986). For any particular size of insect prey, capture and
consumption times are fixed and not under the control of the bat. However, a bat
may be able to change the pursuit portion of handling time. A bat attempting to
intercept prey may be forced to deviate suddenly from its original course to make
the capture because of evasive maneuvers by the prey. This deviation is what is
meant by pursuit.
Moths are an important dietary item of insectivorous bats (Ross, 1967;
Black, 1974). Some moths have ears that are tuned to the echolocation frequency
of bats that prey heavily upon them (Roeder, 1967; Fullard, 1987). This gives
moths the ability to detect and evade bats. It has been demonstrated that such
moths have a 40% survival advantage (Roeder, 1967). Some of the evasive tactics
employed by moths include vertical dives to the ground and tight spirals within
the turning radius of the pursuing bat (personal observation). When insect density
is high, these evasive tactics by moths may mean that a bat will not increase its
net rate of energy gain by pursuing prey it did not capture on the first pass. In
times of high insect density, search time is low and the bat is reasonably certain of
encountering the next prey item soon. Time spent pursuing detected prey might
be better spent attempting to encounter an equivalent prey without pursuit. When insect density is low, it may become profitable to pursue any insect encountered
(within size limitations) since search time is high.
53 This study addresses the predictions of the differential detectability
hypothesis, the differences in prey available in the open and closed habitats, and
the pursuit of prey in the open habitat at different densities of insects.
METHODS
STUDY SITES
The study was carried out at three sites on the island of Hawai \ i known to
be used by bats on a regular basis for foraging (Chapter 2). Two of the sites,
Ocean View and Red Cinder Road, were closed habitat sites. They were located
on the leeward side of the island in the south Kona and Ka \u districts, respectively, both at an elevation of about 600 m. At these sites the bats foraged
below the tree canopy at heights of 1.5 to 15 m above ground, within or near vegetation, and came to within one meter of the foliage. Flight in the closed habitat was slow (6.56 ms") and maneuverable (Chapter 3).
The third site was an open habitat site located in the saddle between two volcanic mountains, Mauna Loa and Mauna Kea, at the Bradshaw Airfield in
Pohakuloa, at an elevation of 1 850 m. Here the bats foraged around a single mercury vapor light attached to the top of a 9 m pole. There was no vegetation higher than 6 m. Bats foraged at a height of 9 m or higher, and very seldom foraged below a height of 6 m. Flight in the open habitat was faster (11.39 rns") and less maneuverable than in the closed habitat (Chapter 3).
54 DIET
The diet of bats was determined by examining: (1) fecal samples collected
from the ground below foraging bats, (2) fecal samples obtained from bats
captured in mist nets, and (3) culled parts of insects collected from the ground beneath foraging bats (at Pohakuloa). Insect parts found in the feces were identified to order using taxonomic keys (Borror, 1981) and a reference collection
of insects, collected at the site during collection of feces. The number of each
insect order in each fecal sample was estimated. The body length of each dietary item was estimated, where possible, by comparing legs, antennae, and wings of insects found in the feces with whole insects in the reference collection.
INSECT FAUNA
The available insect prey was sampled, each night when feces were collected, with a 22 watt battery-operated blacklight trap (BioQuip Products, 1320
E. Franklin Ave., EI Segundo, California). Insects were sorted into orders and body length categories, and counted. The insects were then dried at 95°C for 24 hours before being weighed.
EXPERIMENTS
Three experiments in which the bats were offered artificial prey, tossed into the air, were conducted at Pohakuloa (open habitat). The artificial prey consisted of spheres of children's modeling clay (Play-Doh, Tonka Corporation).
55 Care was taken to throw the spheres to approximately the same height
(approximately 9 m) each time. The top of the 9 m light pole was used to
estimate the height of throws.
One experiment examined size as a criterion for prey selection by bats, and
the extent to which bats in the open habitat at Pohakuloa took different sized
insects than bats in the closed habitats. Bats were offered spheres of three
different sizes corresponding to the range of sizes of insects taken by bats at the
two closed sites. The smallest sphere had a diameter of 5-7 mm which included
the median prey size taken by the bats in the closed habitats. The other two sizes
of spheres were medium (9-11 mm diameter) and large (15-17 mm diameter).
The artificial prey were offered, one at a time, in 9 trials. Each trial consisted of ten spheres of each of the three sizes (total of 270 tosses). Sizes were offered to the bats at random. A positive response was scored for each sphere if bats deviated from their flight paths to chase the sphere as it fell to the ground. A negative response was recorded if no bat deviated from its flight path.
A second experiment involved offering small and large spheres while monitoring any changes in the echolocation calls of the bat in response to the artificial prey. A single trial was conducted in which small (n = 25) and large (n
= 13) spheres were offered to the bats in random order, one at a time. The echolocation calls and the voice of an observer noting the bat's response (as above) to the artificial prey, were recorded on a Racal 4DS instrumentation tape recorder (Racal Recorders, Southampton, United Kingdom) using an ultrasonic
56 microphone. The recordings were then played back and the acoustic responses of
the bats noted as an increase in the pulse rate of the echolocation signal. A
moderate increase indicated that the bat was switching from the search-approach
phase of its echolocation signal to the tracking phase. A marked increase in the
pulse rate indicated that the bat had moved into the terminal, or feeding buzz,
phase of its echolocation signal. This experiment was designed to determine if
smaller sized insects were excluded from the diet of bats because their
detectability decreased as a result of lower frequency echolocation signals used by
the bats.
A third experiment was performed over three nights to determine if bats
varied their handling time of insect prey (decreasing it by not pursuing insects) in
response to changes in prey abundance. Bats were offered spheres of Play-Doh 15
mm in diameter after preliminary tests showed that bats respond to spheres of this
size. The number of large insects visible around the light at the time of the trials,
and the total number of insects trapped on the nights on which trials were
conducted, were also recorded. Each trial consisted of ten spheres, tossed one at a
time. A positive response was recorded for each sphere if a bat deviated from its
original flight path to chase the sphere as it fell to the ground. A negative
response was recorded if no bat deviated from its flight path. This experiment was augmented by comparing the number of pursuits of real prey, on a night
when insect abundance was low, with pursuits on a night when insect abundance was high.
57 RESULTS
DIET IN RELATION TO SIZE AND TAXA
There were more arthropod orders in the two closed habitats (Figures 4.1 and 4.2) than in the open habitat (Figure 4.3). Arthropod prey in the open habitat
(Pohakuloa) consisted of moths, dipterans, and little else.
Although some arthropod taxa were not utilized by L. c. semotus at some sites, no insect order was consistently excluded from its diet (Figures 4.1 to 4.6).
For example, although dipterans were well represented in the insect fauna at Red
Cinder Road and Pohakuloa, they were not taken by the bats (Figures 4.2 and
4.3). At Pohakuloa, the open habitat, only one bat ate anything other than moths during June/July, 1992 (Figure 4.6), while the diet consisted exclusively of moths during September/October, 1992 (n = 5), even though dipterans were abundant.
At Ocean View, on the other hand, bats ate dipterans but did not eat hymenopterans. Hymenopterans were, however, taken at Red Cinder Road.
In contrast, insects in some size ranges were not taken. Insects taken by the bats in the closed habitats were 3 - 12 mm in body length at Red Cinder Road, and 3 - 15 rnm at Ocean View (Table 4.1). In the open habitat, at Pohakuloa, the size range of dietary items was 14 - 24 mm. Thus, insects below 3 nun in body length were excluded from the diet of the bats even though they were often abundant at all three sites (Figures 4.7 to 4.9).
The size ranges of insect orders that were under-utilized by the bats in some habitats suggest that these orders may be excluded because they are too
58 small. For example, most dipterans at Red Cinder Road and Pohakuloa were less than 3 nun in body length and were under-utilized by the bats at these two sites even though they were in relatively great abundance (Figures 4.2 and 4.3).
However, dipterans were taken by the bats at Ocean View where between 10 to
20% of dipterans had body lengths greater than 3 nun (up to 15 mm).
Hymenopterans, almost all of which (95 - 100%) were less than 3 mm in body length, were under utilized by the bats at this site. At Red Cinder Road, where hymenopterans were utilized, 21% of them had body lengths greater than 3 mm.
Small moths were also under-utilized. Although beetles made up a smaller proportion of the insect fauna than moths at the two closed sites (Figures 4.1 and
4.2), there was no difference in the proportions of the diet (taken over alI months) represented by these two orders at the two closed sites (Z = 0.83, 0.2 < p < 0.5; and Z = 0.25, p > 0.5, for Ocean View and Red Cinder Road, respectively, Zar,
1984, pp.395-397). This was probably because most of the available moths (63%) in the closed habitats were less than 3 nun in body length, whereas most beetles
(58%) were between 6 and 12 mm in body length. In the open habitat
(Pohakuloa), 82% of the moths had a body length greater than 14 mm. These results suggest that size influences the diet of L. c. semotus more than taxon, as predicted by the differential detectability hypothesis.
59 DIET VS PREY AVAILABLE IN THE OPEN AND CLOSED HABITATS
There was a difference in the size range of insects taken, and in prey taken relative to prey available, between the two kinds of habitat (Table 4.1). The mean
(and median) size of insect prey taken by the bats in the open habitat was larger than that taken in the closed habitats at Ocean View and Red Cinder Road
(Table 4.1; t = 60.99, p < 0.00001; and t = 67.18, p < 0.00001, respectively). This is consistent with the prediction of the differential detectability hypothesis that the size of insect prey eaten by the bats is larger in the open habitat than in the closed habitats.
This may simply be a reflection of differences in the size range of insects available between the two kinds of habitats. Most (80 - 90%) of the insects in the two closed habitats fell into the three smallest size classes, whereas most of the insects in the open habitat fell into the two largest (53%) and the lowest (33%) size classes (Figures 4.7 to 4.9). However, in the closed habitats bats utilized all available sizes of insects with the exception of those less than 3 mm in body length (Table 4.1, Figures 4.7 and 4.8). In contrast, although there was prey in the size range of those taken by the bats in the closed habitats, bats in the open habitat only utilized prey larger than 14 rnm in body length (Table 4.1, Figure
4.9). This suggests that bats in the open habitat are either ignoring smaller insects or are unable to detect them.
60 DIFFERENTIAL DETECTABILITY
Bats responded more to the large spheres than to the medium or small spheres (Table 4.2; Mann-Whitney, W = 61.5, P < 0.04; W = 125.5,P < 0.0005, respectively). Apart from a single chase in trial 7, bats appeared oblivious to the small spheres. This is supported by the results of the second experiment in which the acoustic responses of the bats were monitored (Table 4.3). There was an acoustic response (an increase in pulse rate) by the bats on only 24% of the small spheres compared to 100% on the large spheres (Z = 2.413, P < 0.02; Zar, 1984, pp.395-397); feeding buzzes on 4% of the small spheres, compared to feeding buzzes on 38.5% of the larger spheres (Z = 0.947, P > 0.05) and chases with no captures on 4% of the small spheres, compared to chases with captures on 61.5% of the large spheres (Z = 2.857, P < 0.01 and Z = 1.675, P < 0.01, respectively).
These results are consistent with the third prediction of the differential detectability hypothesis that bats should respond less to smaller artificial prey than to larger artificial prey when using fast flight in the open habitat.
VARIATION IN HANDLING TIME IN RESPONSE TO CHANGES IN PREY
DENSITY
L. c. semotus appeared to alter its handling time in response to changes in prey density. On the first night when insect density was high, bats did not initially pursue the spheres (trials 1 and 2, Table 4.4). As the night progressed, insect density decreased and the bats chased 50% of the spheres in the third trial. On
61 the second night, insect density was high the entire night and bats did not make
any attempt to catch the spheres in all three trials. On the third night, insect
density was low the entire night and bats responded to the spheres in all three
trials (Table 4.4). These results were supported by the number of pursuits of real
insect prey. There were a greater number of pursuits per 5 minute interval on the
third night (mean = 9±3.22, n = 15) when insect density was low, and no pursuits
per 5 minute interval (mean = 0, n = 15) on the second night when prey density
was high (sign-test, P < 0.0003).
DISCUSSION
The Hawai'ian hoary bat conforms to the predictions of the differential
detectability hypothesis. Fast flying bats appear to select large prey as a result of
constraints imposed upon them by their echolocation system and flight
performance. Large prey are included because they seem to be detected sooner
and more readily than small prey. Studies by Belwood and Fullard (1984) in open
habitat on Kaua'i corroborate the current findings. Bats on Kaua'i used the
same flight speed (11 ms") and the same echolocation call frequency (J. Fullard,
D. S. Jacobs and R. M. R. Barclay, unpublished data) used by the bats in the
open habitat at Pohakuloa. They also fed almost exclusively on large (~ 10 mm body length) moths, even though small flies made up by far the greater proportion
(65.7%) of the insects available to the bats. Nevertheless, there are differences in prey type and size between habitats in which the bats fly slow or fast, and the
62 potential interaction between prey and habitat needs to be considered. In addition, the bats show an ability to alter their foraging behavior in relation to density of prey. Here I discuss issues associated with the differential detectability hypothesis and the variation in foraging behavior.
The key issue surrounding the differential detectability hypothesis is whether or not a fast flying bat perceives the prey it ignores. Failure to respond can occur either if the bat fails to detect a prey item, or if it rejects the prey item while still in the search-approach phase of its echolocation signal. The fact that the bats did respond acoustically and attempted, unsuccessfully, to capture some of the small spheres, suggests that the small spheres, rather than being rejected, were not being detected as readily as the large spheres. This view is supported by the observation that a large (55 mm body length) black witch moth, Otosema odora, was only rejected by bats after physical contact had been made. At least ten capture attempts by bats were observed on this moth. This suggests that the decision to accept or reject a potential prey item is made after capture rather than before. This increases the likelihood that small prey were excluded from the diet because bats could not effectively detect them. Even if bats did detect the small spheres, they may not have done so soon enough, or completely enough for successful tracking and capture.
Variation in foraging behavior under different densities of prey in the open habitat indicates that some aspects of foraging tactics are under the control of the bat even within the constraints imposed upon it by fast flight. Foraging strategies
63 can be prey density-dependent. For example, changes in foraging tactics (Davies,
1977) and search time (Fitzpatrick, 1981), in response to changes in prey density, have been reported in birds. The effect of changes in prey density on prey choice has been considered theoretically in the form of the Prey Model (Charnov, 1976;
Krebs and Davies, 1978; Stephens and Krebs, 1986, and others), and empirically
(Werner and Hall, 1974; Krebs et aL 1977). Although the Prey Model (also referred to as the classical model of prey choice by Krebs and Davies, 1978) treats handling time as a constraint, many foragers can vary their handling time through the partial consumption of prey items (Hassel et al. 1976; Cook and Cockrell,
1978; Giller, 1980; Sih, 1980; MacNair, 1983; Lucas and Grafen, 1985).
Handling time extends from the time prey is detected to the time it is ingested and includes pursuit, capture, and consumption. In examples of partial consumption of prey, it was the consumption component of handling time that was varied. In contrast, the Hawai \ ian hoary bat appears to alter the pursuit component of its handling time of insect prey. There were no pursuits when prey density was high, but many pursuits when prey density was low (Table 4.4). This may enable the bat to increase its rate of energy intake. When prey density is high, search time is low, and the bat is reasonably sure of encountering the next prey item soon. In this situation it may not benefit the bat to spend time chasing prey it cannot capture on the first approach. On the other hand, when prey density is low, search time is high, and the bat can increase its rate of energy
64 intake by chasing insects that it encounters. This appeared to be the strategy employed by the Hawai ' ian bat in the open habitat (Pohakuloa).
The difference in size range of arthropods in open and closed habitats, and the difference in flight behavior in the two habitats, provides no opportunity for comparing the two forms of flight behavior in the same environment. It does not enable one to determine if it is the constraint of insect prey or the constraint of the habitat that influences the foraging behavior of this bat. Differences in the degree of clutter between the two habitats may influence the foraging behavior in that the bat may be constrained to use slow flight in the closed habitats for increased maneuverability for obstacle avoidance. On the other hand, slow, maneuverable flight may also be more efficient for exploiting a prey population in which large prey are rare. If the bat used fast flight in the closed habitat, loss in detectability of smaller prey, as a result of having to use an echolocation call of longer wave-length, would render most of the prey unavailable to it. Slow flight may be more appropriate in the closed habitat because of the increased clutter, and/or because smaller insects were the most common insects in this habitat.
Conversely, bats may be constrained to use fast flight in the open habitat because of lower air density. The Ocean View and Red Cinder Road sites (closed habitats) are at lower elevations than the open habitat at Pohakuloa. The lower air density at Pohakuloa may require the bat to fly faster to generate the same amount of lift. This would also necessitate using a lower frequency echolocation call to increase detection range. Air density can also affect echolocation calls
65 directly (Griffin, 1971). Lower air density would decrease the amount of atmospheric attenuation of sound, and the bat may alter the frequency of its echolocation signal, and therefore also its flight speed, accordingly.
Differences in air density cannot explain the differences in flight speed.
Equating the equations for lift (Norberg 1990, pp.17) at the two elevations reduces them to:
ozVz = GIVI where 02 and Vz are the air density and air velocity (or flight speed) respectively at the higher elevation, and aI and VI the same parameters for the lower elevation. Inserting the actual values for air density at the different elevations
(Humphreys, 1940) gives the required flight speed at the higher elevation as 1.079 times that at the lower elevation. This translates to a speed of 7.08 ms" which is much lower than the actual flight speed (11.39 ms") of L. c. semotus at the higher elevation Pohakuloa site.
Differences in air density also cannot explain the differences in echolocation call frequency. Low frequency sounds are attenuated less by the atmosphere than high frequency sound (Griffin, 1971), and denser air will result in more attenuation. If the degree of attenuation were responsible for differences in echolocation calls, and therefore flight speed, the lower frequency echolocation calls should be used at the lower elevation sites, where air density is higher, rather than at the higher elevation site. Fast flight in the open habitat may thus be related to the fact that large insects were more common than smaller ones. The
66 loss in detectability of smaller insects would not drastically reduce the proportion of the insect fauna available to the bat. Fast flight might also enable the bat to increase its encounter rate.
In conclusion, although a portion of the insect fauna is unavailable to fast flying bats, as a result of constraints imposed upon them by their flight and echolocating abilities, the bats still appear to forage in ways that increase their net rate of energy gain. They are able to respond to changes in insect prey density, within the constraints of their flying and echolocation abilities, by altering their handling time. Handling time in such cases should be treated as a variable under the control of the animal, rather than as a constraint.
67 CHAPTER 5
CHARACfER RELEASE IN THE ENDANGERED HAWAI'IAN HOARY BAT,
LASlURUS ClNEREUS SEMOTUS.
INTRODUCfION
Ecological theory regards interspecific competition as a significant force
that shapes the morphology and behavior of organisms (Darwin, 1950; Lack, 1947;
Brown and Wilson, 1956; Diamond, 1978; Connell, 1983; Grant and Schluter,
1984; Grant, 1986). A corollary to this is that characters shaped by interspecific
competition may change in the absence or reduction of such competition, a
phenomenon known as character release (Connell, 1980). Character release is the
process by which a morphological character state of a species changes under
natural selection arising from the absence, in the same environment, of one or
more species similar to it ecologically and/or reproductively (Grant, 1972). Grant
restricted the term character to morphological attributes of an organism and
considered parallel ecological and behavioral phenomena (e.g., changes in
foraging behavior, habitat, or diet) as examples of ecological release. A change in
ecological niche is necessary for identifying a morphological trait as a potential
case of character release.
The character release hypothesis consists of a number of assumptions and a
major prediction. The three assumptions are: (1) morphological change has
occurred, (2) the morphological change was an adaptive result of natural selection operating on heritable characters, and (3) the morphological change was
68 permitted by the absence or reduction of interspecific competition. The major
prediction of the hypothesis is that some or all aspects of the niche of the
organism have expanded beyond those inherited from its ancestral population to
include those of the putative competitors of the ancestral population in the zone
of sympatry. Such expansion is only possible if permitted by the existence of vacant niches. Vacant niches are defined in terms of the niches occupied by the
putative competitors of the ancestral population.
Evaluating the character release hypothesis is difficult because of reliance upon evidence that the evolution of a population of organisms has been shaped by an absence or reduction of interspecific competition in the past. Since we cannot reconstruct events of the past precisely, character release must be inferred from present events through an evaluation of the assumptions upon which it rests, and the predictions it generates (Grant, 1986, pp.301). This study attempts to do so with respect to the foraging ecology of the Hawai\ ian hoary bat, Lasiurus cinereus semotus.
Character release has been reported in birds (Fjeldsa, 1982; Grant, 1986), while ecological release has been reported in a variety of organisms including birds (Grant, 1965; 1966; Crowell, 1962; Keast, 1968; Diamond, 1970; Fjeldsa,
1982; Grant, 1986), rodents (Cameron, 1964), and lizards (Williams, 1969).
However, in most cases of ecological and character release natural selection has not been demonstrated as being responsible for the release (Grant, 1972). To my knowledge no one has reported character release in bats. This is surprising since
69 bats are the only terrestrial mammals (because of their flying abilities) to have reached isolated oceanic islands (e.g. Hawaivi and the Galapagos), without human help. It is on such isolated islands that the greatest divergence associated with character release would be expected.
The Hawai'ian hoary bat is thought to be derived from the North
American hoary bat, Lasiurus cinereus cinereus (Sanborn and Crespo, 1957;
Tomich, 1986; Fullard, 1987). Both bats are aerial insectivores, i.e., they fly continuously while foraging and catch flying insect prey (Whitaker and Tomich,
1983; Barclay, 1985). The North American hoary bat occurs sympatrically with a number of other insectivorous bats (Whitaker, 1972; Black, 1974; Barclay, 1985), and it appears that partitioning of the habitat, spatially and temporally (Jones,
1965; Kunz, 1973), and partitioning of insect prey, by order and size (Ross, 1967;
Black, 1974), occurs in these bat communities.
Apart from one other bat in the fossil record, the Hawai ' ian bat is the only extant bat in Hawai-I (Tomich, 1986; Ziegler and Howarth, unpub. mans.). It thus appears to have evolved with less interspecific competition than its continental ancestor, which exists in a community with 13 other species of insectivorous bats (Black, 1974). The Hawai-ian bat represents an excellent opportunity for studying possible character release. Previous studies suggest that divergence has occurred in the Hawai\ ian hoary bat: it has undergone a reduction in size (Tomich, 1986) and appears to have a more generalized diet than its North
American counterpart (Whitaker and Tomich, 1983). Whitaker and Tomich (1983)
70 suggest that the more generalized diet of the Hawai\ian bat is due to competitive release.
Character release in a bat can be manifested in its wing and skull morphology, and in characteristics of its echolocation call. Echolocation calls of the Hawai \ ian hoary bat have been investigated (James Fullard, David Jacobs, and Robert Barclay, mans. in prep.) but will be reported under separate cover.
The marked reduction in body size (45% smaller) was associated with allometric responses in the size of its wings (Chapter 3). The decrease in body size per se has marked aerodynamic implications in that it results in a decrease in wing loading and induced power necessary for slow flight. The high aspect ratio of the ancestral species, necessary for fast flight, has been retained. The increased flexibility in flight permitted the expansion of the foraging habitat of the Hawai \ ian bat to include the foraging habitats of the North American hoary bat and that of putative competitors of the North American hoary bat (Chapters 3 and 4).
In contrast, smaller body size was associated with a relative increase in the size of the skull with non-allometric changes in the a:b ratio, size of the masseter muscle, and height of the coronoid process. These changes give the jaws of the
Hawai \ ian hoary bat a relatively larger gape and greater crushing power. These features have been associated with the increased utilization of large and hard bodied insects by insectivorous bats (Freeman, 1979; 1981). Many of the putative competitors of the North American hoary bat, e.g., Myotis evotis, Myotis thysanodes, and Eptesicus fuscus, feed predominantly on hard-bodied insects like
71 beetles (Black, 1974). The jaw structure and musculature of Myotis evotis suggest that it is adapted to consume hard prey (Freeman, 1981).
Here I test the prediction of the character release hypothesis with respect to the diet of the Hawai \ ian hoary bat. The bat should have a more generalized diet in terms of the utilization of hard-bodied prey and the size range of dietary items eaten, as suggested by the divergence in its skull morphology. The validity of the assumptions of the character release hypothesis are also discussed.
METHODS
Data on the size and taxa of arthropods eaten by the Hawai \ian hoary bat were obtained from analysis of fecal samples collected from bats captured in mist nets and from the ground below foraging bats (Chapter 4). Collections were made in both open and closed habitats (Chapter 4). These data were compared with published data for the North American hoary bat (Ross, 1967; Black, 1974;
Barclay, 1985; 1986; Warner, 1985). Diets of the two subspecies were compared in terms of the size of insects taken, and the relative proportions of moths and beetles eaten. Size has been shown to be an important characteristic of insect prey for both subspecies, and perhaps bats in general (Chapter 4, this study; Barclay
1985; Barclay and Brigham, 1991). Black (1974) divided North American insectivorous bats into species that fed predominantly on moths, which he called
"moth strategists", and species which fed predominantly on beetles, which he called "beetle strategists". Because moths and beetles appear to be important
72 prey items in the diets of insectivorous bats, their relative proportions in the diets of bats were used for interspecific comparisons.
RESULTS
SIZE RANGE OF DIETARY ITEMS
The sizes of insects taken by the Hawai \ ian bat in the open habitats were
15-23 mm in body length. This is similar to the size range taken bythe North
American hoary bat which forages in open habitats (Figure 5.1). The size of dietary items taken by the North American hoary bat included beetles with body lengths of 15 mm and moths with body lengths of 14-29 mm (Barclay, 1985).
The size range of insects taken by the Hawai>ian hoary bat in the closed habitats were 3-15 mm in body length. This is similar to that taken by putative competitors of the North American hoary bat, the silver-haired bat, Lasionycteris noctivagans, and the big brown bat, Eptesicus fuscus. The sizes of insects taken by
L. noctivagans ranged from 2-10 mm in body length and from 3-10 rnm in wing length (Barclay, 1985), and that taken by E. fuscus ranged from 6-12 mm (Black,
1974) in body length (Figure 5.1).
The size range of insects taken by the Hawai>ian hoary bat thus included the size range of the North American hoary bat and that of the putative competitors of the North American hoary bat. With the exception of very small insects (less than 3 mm body length), the size range of insects taken by the
Hawaivian hoary bat reflects the size range of insects available (Chapter 4,
73 Figures 4.7, 4.8, and 4.9). The North American hoary bat, on the other hand, only takes insects in the upper size range of those available, while its putative competitors only take those in the middle to lower range of the sizes available
(Figure 5.1; Black, 1974; Barclay, 1985).
BEETLES VS MOTHS
The expansion of the diet of the Hawai \ ian hoary bat, that includes the diets of the North American hoary bat and its putative competitors, is also reflected in the proportions of moths and beetles it utilizes (Figure 5.2). In the open habitat the Hawai \ ian hoary bat does what its North American counterpart does; it feeds predominantly on large moths. In the closed habitats its diet consists predominantly of beetles, similar to the diet of the putative competitors of the
North American hoary bat, E. fuscus, Myotis thysanodes, and Myotis evotis (Black,
1974).
The proportions of beetles and moths in the diet of the Hawai \ ian hoary bat generally reflect the availability of these insects in the insect fauna (Chapter
4). However, this was not true for the North American hoary bat. Although there are enough beetles to support at least three other species of bat (E. fuscus, M. thysanodes, and M. evotis) the North American hoary bat fed predominantly on moths (Black, 1974). Furthermore, at Manitoba, Canada, beetles made up the greatest proportion of the diet of the North American hoary bat in June, when overall insect abundance was low, but a relatively minor proportion of the diet in
74 July and August, when overall insect abundance was high (Barclay, 1985). This
occurred even though the relative abundance of beetles was higher in July and
August than in June. Seasonal differences in the diet of the North American
hoary bat suggest that beetles are utilized to a significant extent by the North
American hoary bat only when overall insect abundance, especially that of soft bodied insects, is low.
DISCUSSION
The diet of the Hawai \ ian hoary bat has expanded, in terms of size and
taxa of insects eaten (Figures 5.1 and 5.2, respectively), to include the diets of the
North American hoary bat and its putative competitors. There is a close correspondence between the morphological divergence of the Hawai \ ian bat and its diet. The dietary expansion was associated with increases in the relative size of the skull, the gape of the jaws, the size of the masseter muscle, and the height of the coronoid process, giving the jaw more crushing power for more efficient processing of large and hard-shelled prey (Chapter 3). These results are consistent with the prediction of the character release hypothesis. Here I evaluate the character release hypothesis as an explanation for the divergence in the
Hawai \ ian hoary bat.
The divergence in the skull morphology and diet of the Hawi \ ian hoary bat is similar to what Grant (1986) found in Darwin's finches, where a change in feeding structure was associated with changes in the hardness and size of dietary
75 items. The difference here is that whereas Darwin's finches can simply walk to a seed, the Hawai\ ian bat has to fly after its insect prey. Changes in feeding
structure and diet have also been accompanied by changes in the flight
morphology of the Hawai \ ian hoary bat. The decrease in wing loading permits
the slow, maneuverable flight necessary for the exploitation of closed habitats, which is where the bat takes all of the beetles in its diet (Figure 20). The
Hawai \ ian bat has also retained the high aspect ratio of the ancestral species
which permits efficient and rapid flight in open habitats. These changes have
permitted the Hawai \ ian bat to expand into the foraging niches of the putative
competitors of the North American bat, while retaining the ability to exploit the
ancestral niche, as predicted by the character release hypothesis.
I have documented divergence in the Hawai \ian hoary bat (Chapter 3) and
showed here that this divergence is consistent with the prediction of the character
release hypothesis. However, problems are inevitably encountered when one tries
to infer a historical phenomenon like character release from an analysis of present
events. Enough may not be known, or knowable, about the history of the derived
and ancestral populations, or their respective environments. The confirmation of
the prediction of the character release hypothesis is complicated by the fact that,
in addition to eating native moths and beetles, the Hawai\ ian hoary bat also eats
a variety of non-native insects, including moths, beetles, and termites (Chapter 4,
Figures 4.4, 4.5, 4.6). It is not known to what extent the present diet of the
Hawai \ ian hoary bat reflects the historical one, or to what extent the present
76 insect fauna reflects the one to which the Hawai \ ian bat was exposed over its
evolutionary history. Such information may not be obtainable.
Similarly, an implicit assumption made in this study is that the absence of
other insectivorous bats implies the reduction of interspecific competition.
However, interspecific competition can come from other organisms, especially
native insectivorous birds. Competition between unrelated organisms for the same
resource has been demonstrated (Brown and Davidson, 1976; Schluter, 1986).
Furthermore, although none of the extant insectivorous forest birds are nocturnal
or aerial insectivores (Banko and Banko, 1976; Muller-Dombois et. aI., 1981;
Mountainspring and Scott, 1985), and may have only provided reduced or indirect competition, the nature or degree of interspecific competition over the evolutionary history of the Hawai\ ian hoary bat is not known.
This problem is further compounded by the historical existence of at least one other species of bat in Hawai\i. Recent fossil evidence suggests that this extinct bat was contemporaneous with the Hawai \ian hoary bat possibly as recently as one hundred to two hundred years ago (Frank Howarth, Bishop
Museum, pers. com.). The extinct bat has been described as a smaller bat (Ziegler and Howarth, unpub. mans.) which means that it may have foraged in the closed habitat. If so, it may have been in competition with the Hawai \ ian hoary bat.
Although no fossils of this bat have been found on the island of Hawai \ i, it casts some doubt upon the validity of the assumption that interspecific competition has been reduced over the evolutionary history of the Hawai \ ian hoary bat.
77 The assumption that the divergence was the result of natural selection is
also confounded by the required historical perspective. The adaptiveness of a
character suggests that it is the result of natural selection (Futuyma, 1986).
Adaptation is manifested by design to solve a biological problem (Williams, 1966),
and by the comparative method, which views parallel and convergent evolution as
evidence of adaptation (Darwin, 1950; Futuyma, 1986, pp.251-253, Pagel and
Harvey, 1991). The functional significance of the divergence in the Hawaivian hoary bat is indicated by the increase in crushing power of its jaws which has been associated with an increase in the utilization of hard-bodied insects. The
Hawai 'ian bat also has the flight morphology predicted by aerodynamic principles to fly in both open and closed habitats. In addition, the intermediate morphology of the Hawai> ian bat in relation to bats that forage in closed habitats only, and bats that forage in open habitats only, suggests that the correlation between the morphology and foraging ecology of the Hawai' ian hoary bat is consistent with that of other insectivorous bats (Figure 3.8, Chapter 3). There also appears to be evidence of convergent evolution between birds and bats with respect to morphology and habitat use (Norberg, 1986; Norberg and Rayner, 1987).
Although these changes appear to be adaptive, they do not necessarily have to be the product of natural selection (Lewontin and Gould, 1979). They may have been the product of stochastic processes such as genetic drift and founder effect. The changes may have been the result of a single founder event or a number of founder events as each island was colonized in turn. It is also possible
78 that exposure to hard-bodied insects over the developmental history of the bats may have produced the changes in skull morphology. The adaptiveness of these traits could thus be purely incidental. Even if the differences could be interpreted as adaptations, they may be better viewed as exaptations since they may not have been designed by natural selection for their current function (Gould and Vrba,
1982). This is especially true for the wing morphology and associated flight behavior. Divergence in the wing morphology could have been the incidental by product of selection for reduced body size to reduce resource requirements in a resource-limited environment (Sondaar, 1977; Marshall and Corruccini, 1978;
Lomolino, 1985). It may have had little to do with conferring the ability to use both open and closed habitats.
Lastly, and perhaps more importantly, the character state of the Hawai \ ian bat when it arrived in these islands is not known. One is forced, in the absence of adequately dated fossils, to assume that it was the same as that of the modem
North American hoary bat. In doing so, one is also assuming that the North
American hoary bat has not evolved since the establishment of the Hawai\ ian hoary bat population. These may not be valid assumptions.
In conclusion, the Hawai \ ian hoary bat has undergone a marked divergence in its foraging ecology and associated morphology. From an ancestor that was relatively specialized in its flight behavior, habitat use, and diet, the
Hawai \ ian hoary bat has evolved a more generalized ecology. It has developed the ability to use a wider range of habitats and expanded its diet. This increased
79 generalization may have enabled the Hawai \ ian hoary bat to survive in an
environment in which resources may have been limiting while other species of
bats, both endemic and introduced (Tomich, 1986) failed to persist. The bat is a good candidate for character release based on present-day study, but uncertainties remain about the context of the evolution of the divergence.
80 Table 2.1 Numbers of bats at each foraging site on the island of Hawai \ i. Location Percentage of Number Number visits bats were of visits of bats observed Kipahoehoe Natural Area 100 5 10 Reserve (Highway 11, 90-92 mile markers) Manuka Natural Area Reserve 100 5 4 (Highway 11, 80-82 mile markers) Red Cinder Road (off 77 13 5 Highway 11, near 70 mile marker) Highway 11 (70 mile marker) 60 5 7 - 10 Kaimu Bay (Puna) 100 2 3-4 Kalapana (Puna - over new 100 2 5 lava flow) Kalopa State Park (Hamakua) 100 2 2 Honoka> a (St. George's 100 5 2-3 Cemetery, Lehua Street) Waimanu Bay (North Kohala) 100 2 5 Keokea Bay (North Kohala) 100 2 11 Ocean View (Intersection of 100 6 4 Hukilau and Tree Fern Avenues, South Kona) Pohakuloa (Bradshaw Military 100 7 7 - 10 Airfield) Waiono Ranch (Kailua- 100 4 3-5 Kona) Volcanoes National Park 100 2 2-3
Note.--Data are results of surveys conducted between August and October (see text)
81 Table 2.2.
Numbers" of bats at various foraging sites for different months.
Month" Location February May June August September October Kipahoehoe 0.5±0.5 0 1 1O±0.0 Natural (2) (2) (1) (2) Area Reserve Manuka 0.5±0.5 HO.O 0 4±0.0 Natural (2) (2) (1) (2) Area Reserve Red Cinder 0.5±0.5 0 0 5±0.0 3.6±2.2 Road (2) (2) (1) (2) (11) Highway 11 0 0 0 5.7±2.6 (70 mile (2) (2) (1) (2) marker) Kaimu Bay 0 0 0 3.5±0.5 (2) (2) (1) (2) Kalopa 0 0 0 2±0.0 State Park (2) (2) (1) (2) Keokea Bay 0 0 0 11±0.0 (2) (2) (1) (2)
Ocean View O.25±0.5 3.4±0.5 3.4±O.s 5.6±2.5 4.5±1.7 3.5±1.7 (4) (5) (2) (4) (4) (2)
Pohakuloa 0.75±0.5 0.2±0.5 0.7±1.0 0.5±0.6 2.7±2.1 (MPRC, (4) (4) (4) (4) (4) HHP, 1992y
3 Mean±SD, sample size in parentheses. b See text for year. e Multipurpose range complex, data from the Hawai'ian Heritage Program (see list of literature cited)
82 Table 2.3.
Elevation and vegetation associated with each bat foraging site. Location Elevation (m) Vegetation Kipahoehoe Natural Area 386 0'hia lowland mesic N Reserve forest Manuka Natural Area 548 0'hia lowland mesic N Reserve forest Red Cinder Road 593 0'hia lowland mesic M forest/eucalyptus Highway 11 (70 mile 593 Eucalyptus trees/exotic E marker) shrubs Kaimu Bay Sea level Open ocean Kalapana 60 Open lava Kalopa State Park 549 o' hia/eucalyptus forest M
Honoka> a 354 Eucalyptus/macadamia E orchard Waimanu Bay Sea level Open ocean Keokea Bay Sea level Open ocean Ocean view 665 o 'hia lowland mesic N forest Pohakuloa 1830 Mamani-naio (Sophora N chrysophylla and Myoporum sandwicense) open woodland Waiono Ranch 1070 o 'hia forest N interspersed with grass fields Volcanoes National Park 1220 Koa (Acacia koa) forest N
N - Native E = Exotic M = Mixed (native and exotic)
83 Table 3.1.
Body mass' and forearm length' of Lasiurus cinereus semotus and L. c. cinereus. L. c. semotus L. c. cinereus Variable Females Males Females Males Body mass 17.9±2.1 14.2±1.2 32.1±4.1 24.2±2.0 (g) (50) (25) (26) (13) Forearm 50.5±1.1 48.5±1.2 55.1±1.6 51.9±1.5 length (50) (25) (26) (13) (mm)
Mean±SD, sample sizes in parentheses.
84 Table 3.2. Skull dimensionsf (rom) of Lasiurus cinereus semotus and L. c. cinereus. L. c. semotus L. c. cinereus Female Male Female Male (12) (7) (11) (9) Mass 18.0±2.0 16.5±3.3 34.8±3.5 23.8±1.9 Skull length 15.7±O.3 15.0±O.6 16.7±O.4 16.0±O.6 Dentary length 12A±O.3 11.7±O.4 13.2±O.3 12.5±O.3 Dentary thickness 1.6±O.1 1.6±O.1 1.8±O.1 1.7±O.1 Joint to origin of 5.9±O.2 5.6±O.3 6.5±O.2 6.0±O.2 masseter (a) Joint to insertion of 2.8±O.1 2.7±O.2 3.3±O.1 3.1±O.1 masseter (b) a:b* 2.1±O.1 2.1±O.1 2.0±O.1 1.9±O.04 Length of masseter 4.2±O.3 4.0±O.2 4.0±O.1 3.7±O.2 muscle scar* Height of condyle 2.2±O.2 2.l±O.3 2.5±O.1 2A±O.1 Height of coronoid 2.9±O.2 2.8±O.3 2.8±O.1 2.6±O.1 process* Area of maxillary 5.5±OA 5.0±OA 6.3±O.3 5.5±O.3 tooth row (PM4- M3) Length of maxillary 4.8±O.2 4.7±O.2 5A±O.1 5.0±O.1 tooth row (PM3- M3) Height of upper 3.0±O.2 2.8±O.3 3A±O.1 3.1±O.2 canine Maximum width of 1.2±O.1 1.1±O.1 1.3±O.1 1.2±O.O5 upper canine
& Mean±SD, sample sizes in parentheses. * These measurements are larger in L. c. semotus than in L. c. cinereus.
85 Table 3.3.
Allometric exponents of skull measurements regressed against skull length.
Geometric Power r similarity Function. Skull length MO.33 1.218Mo.089 0.902 (±0.018)
Dentary length S1.00 0.824S 0.906 0.892 (±0.207) Dentary thickness S 1.00 0.094S 1.249 0.601 (±0.751) Height of condyle S 1.00 0.109S1578 0.626 (±0.885)
Area of maxillary S2.00 0.026S 1.666 0.675 tooth row (PM4-m3) (±0.821) Length of maxillary S 1.00 0.258S 1.419 0.817 tooth row (PM3-M3) (±0.432) Height of upper S 1.00 0.160S 1.435 0.631 camne (±0.797) Maximum width of S 1.00 0.060S 1545 0.545 upper canine (±1.074)
* Slopes different from zero (P < 0.001), 95% confidence limits in parentheses. M Body mass. S Skull length. r Correlation coefficient.
86 Table 3.4. Wing parameters" for Lasiurus cinereus semotus and L. c. cinereus. L. c. semotus L. c. cinereus Parameter Female Male Female Male Body mass O.O17±O.OO2 O.0l5±O.OOl O.O31±O.OO3 O.O27±O.O04 (kg) (24) (11) (11) (4) Wing span O.351±O.013 O.342±O.Oll O.380±O.O21 O.363±O.O20 (m) (24) (11) (11) (4) Wing area O.O174±O.OOO9 O.O162±O.OOl O.O209±O.OO19 O.0194±O.O019 (nr') (24) (11) (11) (4) Wing 9.92±1.22 9.03±O.46 14.77±1.79 12.13±1.274 loading (24) (11) (11) (4) (Nm-2)
Mean±SD, sample sizes in parentheses.
87 Table 3.5 Allometric exponents of wing parameters regressed against body mass. n Wingspan Wing area Wing loading Geometric M 0.333 M 0.667 M 0.333 similarity L. c. semotus 24 0.431M 0.050 0.028M 0.117 351.7M 0.882· r = 0.170 r = 0.281 r = 0.912
L. c. cinereus 11 1.398M 0.375" 0.040M 0.186 275.2M 0.843·· r = 0.649 r = 0.187 r = 0.663
Populations 35 O.595M 0.130· 0.056M 0.286· 183.5M 0.723· combined (±O.049) (±0.083) (±0.081) r = 0.681 r = 0.773 r = 0.953
* Slope different from zero (P < 0.0001), 95% confidence limits given in parentheses. ** Slope different from zero (P < 0.03). M Body mass. r Correlation coefficient.
88 Table 3.6
Aspect ratio (A), tip area ratio (Ts)' tip length ratio (T1), and tip shape index (I) of Lasiurus cinereus semotus, and L. c. cinereus. L. c. semotus. L. c. cinereus. Parameter Female Male Female Male A 7.13±0.29 7.22±O.35 6.95±O.69 6.82±O.62 (24) (11) (11) (4)
1; 1.37±O.O8 1.32±O.O8 1.52±O.17 1.52±O.25 (24) (11) (11) (4)
t; O.84±O.O6 O.81±O.O5 O.86±O.O7 O.79±O.13 (24) (11) (11) (4)
I 1.60±O.26 1.62±O.30 1.50±O.93 1.lO±O.16 (24) (11) (11) (4)
Mean±SD, sample sizes in parentheses.
89 Table 4.1.
Descriptive statistics for the size of dietary items.
Site N Mean" SD Median" Range" Red Cinder 64 6.07 0.67 5.95 3 - 12 Road3 Ocean 202 6.48 1.58 5.95 3 - 15 View" Pohakuloa" 59 18.92 1.27 19.00 14 - 24
3 Open habitat. b Closed habitat. C Units of measurement are mm.
90 Table 4.2.
Response to artificial prey by Lasiurus cinereus semotus. Trial Small' Medium' Large' (5-7mm diam.) (9-11mrn diam.) (15-17mm diam.) 1 0 20 60 2 0 30 40 3 0 10 10 4 0 20 20 5 0 20 50 6 0 20 30 7 10 40 50 8 0 20 40 9 0 10 30
Percentage of spheres pursued (see text).
91 Table 4.3.
Acoustic responses of Lasiurus cinereus semotus to artificial prey. Small" Large " Acoustic Feedin Chases Acoustic Feeding Chases respons g response buzzes and e buzzes captures 24 4 4 100 38.5 61.5
Percentage of spheres (see text).
92 Table 4.4.
Capture" attempts by bats on artificial prey," Insect density" Trial 1 Trial 2 Trial 3 First > 10 (trial 1 and 2) 0 0 50 night < 10 (trial 3) Second >10 0 0 0 night Third <10 60 40 10 night a Percentage of spheres (see text). b Sphere size used was 15-17mm diameter (see text). C See text.
93 LEGEND Echo Location Survey o No Bats • Bats Seen I:i Bats Heard
N t
10, , I 10 20
Figure 21. Distribution of Hawaii hoary bat sightings and echolocation detections on the island of Hawai'i. The circles indicate 26 sites surveyed for bats; open circles indicate sites at which no bats were seen; closed circles indicate sites where bats were seen on at least two consecutive nights. The dashed lines indicate the roads along which echolocation surveys were conducted; the triangles indicate areas where bats were heard during the echolocation surveys.
94 1.2.- I I "16 [] Insects
tI) Bats 1.0 I- IMY L::',I I I ~5 ..... c C) Q) Q) II) en -II) 0.8 4 rn as ..as ::E m >. ... 0 c 3 - 0.6 ...Q) ..o .Q Q) E ~ II) :::I C z .,.." c c ~·:.:t·" 2 as 04 as Q) :~·:r~ Q) :{.:::: ' ::E ::E .:,.'.: IT . T :·;\:i':. 0.2 1-1:!S'i:1 r,:{:',P~d fi\'!,J)~{;~il w;'::ml --WZt 1.~?:iif6%J 1:,,',Y'li~f31 ~-I1 .",:.:'/{ f.:;',:r: ;~~1~:/ .:~,~.y' 0.0 L fc..:·.:lt;m ,.i'"j"·.;>';U"·S' If.·.·; ··t~!'~'·"j'1 ,"",.-., 'Hz1M' '--~lo;r.rn~ ,. ~i"-~IA\f+1 1'~""";'''''~5i' '::;:·I"\1g.l .J 0 Feb Mar May Jun Jul Aug Sep Oct
Figure 22 Mean insect biomass and mean numbers of the Hawai'ian hoary bat, Lasiurus cinereus semotus, seen at Ocean View on the island of Hawaii, in different months of 1992 WING DESIGN FOR CLOSED HABITATS - MANEUVERABILITY
Low Wing Loading
Low Aspect Ratio
Rounded Wingtips
~
WING DESIGN FOR OPEN HABITATS - SPEED
High Wing Loading
High Aspect Ratio
Pointed Wingtips
Figure 3.1. Morphological characteristics of wings of bats that specialize foraging in closed and open habitats. Figure 3.2a. Measurements taken from skulls of the Hawai'ian hoary bat, Lasiurus cinereus semotus (shown in photograph), and the North American hoary bat, Lasiurus cinereus cinereus; d - height of the condyle (top of the left condyle to the plane of the alveoli of the left first and second molar); e - height of the coronoid process (top of the left coronoid process to plane of the alveoli of the left first and second molar); f _ length of the skull (from the occipital to the alveolus of the canine); i - dentary thickness (from the plane of the alveoli of the left first and second molar to the bottom of the left dentary). The scale line represents 2.5 mID.
97 Figure 3.2b. Measurements taken from skulls of the Hawai'ian hoary bat, Lasiurus cinereus semotus (shown in photograph) and the North American hoary bat, Lasiurus cinereus cinereus; a - distance from the anterior surface of the mandibular fossa to the origin of the masseter muscle (fusion line of the zygomatic arch with the maxilla); b _ distance from the top of the left condyle to the insertion of the masseter muscle (bottom of the left angular process); f - length of the skull (from the occipital to the alveolus of the left canine); g - length 'of dentary (from the back of the left condyle to the epiphysis of the dentary). The scale line represents 2.5 mm.
98 Figure 3.2c. Measurements taken from the skulls of the Hawai'ian hoary bat, Lasiurus cinereus semotus (shown in photograph) and the North American hoary bat, Lasiurus cinereus cinereus; c - length of masseter muscle scar (from the back of the left condyle to the furthest extent of the scar on the left dentary); h - length of left maxillary tooth row (from the front of the left fourth premolar to the back of the left third molar). The scale line represents 2.5 mm.
99 b
Ihw law
-8
Wing loading B = Mg/S
Aspect ratio A = b2/S
Tip shape index I = TsI(T. - Ts)
where T; = Tip area ratio = Shw/Saw
T1 = Tip length ratio = lhw/1aw
Figure 3.3. Morphological parameters used to describe bat wings (after Norberg and Rayner, 1987). ~ *~ -E 2f,X Z - ** ",* () x -.c.... 1.5 eCI .!! -:3 ~ JI:: 0.097 0.33 o CI) ~ M (M] z L. c. semotus * L. c. clnereus 1 ::.-' --1.., i, , 10 20 30 Body mass (g)
Figure 3.4. Skull length versus body mass for the Hawai'ian hoary bat, Lasiurus cinereus semotus, and the North American hoary bat, L. c. cinereus. Both axes have logarithmic scales (base 10). The line is the least squares regression line. M- and [Mol are mass raised to the calculated exponent and the expected exponent for geometric similarity, respectively. 0.5 .-
-e 0.4 ~*-* e z z ** -eu l* C i x O) * Cl * it:: 0. 3 .- 0.13 ~ [M 0.33) t3 M z l. c. semotus * l. c. cinereus
0.02 0.03 Body mass (kg)
Figure 3.5. Wingspan versus body mass for the Hawai'ian hoary bat, Lasiurus cinereus semotus, and the North American hoary bat, L. c. cinereus. Both axes have logarithmic scales (base 10). The line is the least squares regression line, M- and [MOl are mass raised to the calculated exponent and the expected exponent for geometric similarity, respectively. 0.03
CII- E -eu Q) 0.02 ~ z eu x C) c ~ * ~ 0.286 0.67 S I x L. c. semotus M(M] L. c. cinereus I * 0.01 0.01 0.02 0.03 Mass (kg)
Figure 3.6. Wing area versus body mass for the Hawai'ian hoary bat, Lasiurus cinereus semotus, and the North American hoary bat, L. c. cinereus. Both axes have logarithmic scales (base 10). The line is the least squares regression line. M- and [M-] are mass raised to the calculated exponent and the expected exponent for geometric similarity, respectively. 100 L I
CII I- zE -Cl .5 1.0 * "D ~z * ell 0 Cl 0.723 C 0.33 i M (M I -~ z L. c. semotus * L. c. cinereus 1 0.01 0.02 0.03 0.04 Body mass (kg)
Figure 3.7. Wing loading versus body mass for the Hawai'ian hoary bat, Lasiurus cinereus semotus, and the North American hoary bat, L. c. cinereus. Both axes have logarithmic scales (base 10). The line is the least squares regression line. Mo and [MOl are mass raised to the calculated exponent and the expected exponent for geometric similarity, respectively. 100 I I
CII I- E z _ ...e---g - --- -Cl 10 c + -----* '0as - 0 o E. fuscus Cl Z L c. semotus 0.723 0.33 c M [M) i + L noctlvagans * L. c. clnereus 1 ~ 0 0.01 0.02 0.03 0.04 Vl Body mass (kg)
Figure 3.8. Mean wing loading plotted on logarithmic coordinates against mean body mass fur the Hawai'ian hoary bat, L. c. semotus, the North American hoary bat, L. c. cinereus, the silver-haired bat, Lasionycteris noctivagans, and the big brown bat, Eptesicus fuscus, in relation to the regression line of wing loading versus body mass for the two hoary bat subspecies. M" and [M'] are mass raised to the calculated exponent and the expected exponent for geometric similarity, respectively. Col Lep ~ Iso ... Hem ~ (I) - "C o Hom I +00 0 ~ Q) Pso tJ) n 1:1 1 c: Dip n • 10 ~
~ Hym ~ ~ Ort I Unk I I 100 80 60 40 20 0 20 40 60 80 100 Diet (Ofo) Trapped (Ofo) August 1991 (a)
Figure 4.1a. Occurrence of insect orders trapped (% of total trapped) and in the diet (% of total insects occurring in all fecal samples) of the Hawai'ian hoary bat, Lasiurus cinereus semotus, at Ocean View (closed habitat) on the island of Hawai'i, August 1991; Col - Coleoptera; Lep - Lepidoptera; Iso - Isoptera; Hem - Hemiptera; Hom Homoptera; Pso - Psocoptera; Dip - Diptera; Hym - Hymenoptera; Ort - Orthoptera; Unk - Unknown. Col Lep Iso '- Hem Q) ~ "C o Hom . ~ +00 0 Q) Pso I tn n II 11 n II 4 c: Dip ~ .... Hym ~ 0 -..) Ort I • Unk I I I I I J I I 100 80 60 40 20 0 20 40 60 80 100 Diet (%) Trapped (%) Sept/Oct 1991 (b)
Figure 4.1b. Occurrence of insect orders trapped (% of total trapped) and in the diet (% of total insects occurring in all fecal samples) of the Hawai'ian hoary bat, Lasiurus cinereus semotus, at Ocean View (closed habitat) on the island of Hawai'i, September/October 1991. Abbreviations are the same as in Figure 4.1a. Col I ~ Lep I .~ Iso '- ~ Hem , '- I 0 .... Hom 0 m Q)
~ ~ n rJ 4 - Pso n • 9 Dip ~ I 0 00 Hym ~ Ort I I I I I -..., I I 100 80 60 40 20 0 20 40 60 80 100 Diet (Ofo) Trapped (Ofo) June 1992 (c)
Figure 4.1c. Occurrence of insect orders trapped (% of total trapped) and in the diet (% of total insects occurring in all fecal samples) of the Hawai'ian hoary bat, Lasiurus cinereus semotus, at Ocean View (closed habitat) on the island of Hawai'i, June 1992 Abbreviations are the same as in Figure 4.1a. Col I ~ Lep -~
i L- Q) Iso 'C aL- .... Hem ~ 0 Q) en n a 11 n a 5 c: Hom ~ ..... ~ Dip I ~
Hym I ~ I I I I I I 100 80 60 40 20 0 20 40 60 80 100 Diet (Ofo) Trapped (Ofo) Aug/Sep 1992 (d)
Figure 4.1d. Occurrence of insect orders trapped (% of total trapped) and in the diet (% of total insects occurring in all fecal samples) of the Hawai'ian hoary bat, Lasiurus cinereus semotus, at Ocean View (closed habitat) on the island of Hawai'i, August/September 1992 Abbreviations are the same as in Figure 4.1a. Col I I~ Lep I I~
1 Iso , ~ Hem Q) E "C 5 Hom ~ ..... 0 Q) Pso n • 5 ~ n • 3 UJ c Dip ~ .... 0 Hym ~ • Unk I I I I 100 80 60 40 20 0 20 40 60 80 100 Diet (Ofo) Trapped (Ofo)
Figure 4.2 Occurrence of insect orders trapped (% of total trapped) and in the diet (% of total insects occurring in all fecal samples) of the Hawai'ian hoary bat, Lasiurus cinereus semotus, at Red Cinder Road (closed habitat), on the island of Hawai'i, September, 1991. Abbreviations are the same as in Figure 4.1a. Col I I Lep I
Hem .~ I Q) "C ~ Pso 0 I ....o Q) Dip n • 28 n • 4 UJ ~ c: .... - .... Hym I r Neu I I I I r I I I 100 80 60 40 20 0 20 40 60 80 100 Diet (%) Trapped (%) June 1992 (a)
Figure 4.3a. Occurrence of insect orders trapped (% of total trapped) and in the diet (% of total insects occurring in all fecal samples) of the Hawai'ian hoary bat, Lasiurus cinereus semotus, at Pohakuloa (open habitat) on the island of Hawai'i, June 1992; Neu - Neuroptera, The other abbreviations are the same as in Figure 4.1a. Col I ~
Lep ... (1) "C... Iso 0 ..... • ~ Hem rn I c n • 29 n • 2 .... Pso I N Dip I ~ I I I 100 80 60 40 20 0 20 40 60 80 100 Diet (Ofo) Trapped (Ofo) July 1992 (b)
Figure 4.3b. Occurrence of insect orders trapped (% of total trapped) and in the diet (% of total insects occurring in all fecal samples) of the Hawai'ian hoary bat, Lasiurus cinereus semotus, at Pohakuloa (open habitat) on the island of Hawai'i, July 1992 Abbreviations are the same as in Figure 4.1a. Lep
"- ,(1) Dip I ~ 'C... 0
~ (J Q) n II 14 I n II 4 t/J Pso I -c:
lJ,) - Hym I I I I I I I I I 100 80 60 40 20 0 20 40 60 80 100 Diet (%) Trapped (%) Sep/Oct 1992 (c)
Figure 4.3c. Frequency occurrence of insect orders trapped (% of total trapped) and in the diet (% of total insects occurring in all fecal samples) of the Hawai'ian hoary bat, Lasiurus cinereus semotus, at Pohakuloa (open habitat) on the island of Hawai'i, September/October 1992 Abbreviations are the same as in Figure 4.1a. 100, I
n • 9 80
60 -'#. -en ....ca m 40
/o--ol /o--ol 20 ,J:..
oI V///l V///I v/co vcL/) V/LLJ V///l V///I I Col Lep Iso Hem Hom Pso Dip Hym Insect order August 1991 (a)
Figure 4.4a. Proportion of bats eating each insect order at Ocean View (closed habitat), on the island of Hawai'i, August, 1991. Abbreviations are the same as in Figure 4.1a. 100, 177?A I
n • 11 80
-?fl 60 -.....en n1 m 40
..... 20 IJt
oI VA/A ra/J VA//1 l'lY/J V/V/I V/V11 V/V/I V/VI1 V/V/I Col Lap Iso Hem Hom Dip Hym Ort Unk Insect order Sept/Oct 1991 (b)
Figure 4.4b. Proportion of bats eating each insect order at Ocean View (closed habitat), on the island of Hawai'i, September/October, 1991. Abbreviations are the same as in Figure 4.1a. 1001 I
n a 8 80
-'#. 60 .-...,en eu m 40
~ ~ 20 0'1
oI V//L/l [////11 [///0 1/////1 (///0 [////<1 v////J I Col Lep Iso Hem Hom Dip Crt Insect order June 1992 (c)
Figure 4.4c. Proportion of bats eating each insect order at Ocean View (closed habitat), on the island of Hawai'i, June 1992 Abbreviations are the same as in Figure 4.1a. 100 'r------n • 11 80
,?ft- 60 -..U) &l 40
20 .... -..J
0 Col Lep Iso Hem Hom Dip Crt Insect order Aug/Sept 1992 (d)
Figure 4.4<1. Proportion of bats eating each insect order at Ocean View (closed habitat), on the island of Hawai'I, August/September 1992 Abbreviations are the same as in Figure 4.1a. 100 .------.------17771 . UM I
• October 1990 - n • 7
80" ,/;' ~' LJ September 1991 - n • 5 »:/X / / , r.// " ~ 60 -0
-CI) ~m m 40 .-
...... 20 .- 00
o ~- ,,,,II_II Lep Iso Hem Hom Dip Unk Insect order
Figure 4.5. Proportion of bats eating each insect order at Red Cinder Road (closed habitat). on the island of Hawai'i. Abbreviations are the same as in Figure 4.1a. 100 I v7777/1 I
n • 6 80
60 -?ft -en ~ m 40
20 -\0
o I V//Y,//1 V//V/Il V//V//1 I Lep Iso Hem- Insect order June/July 1992
Figure 4.6. Proportion of bats eating each insect order at Pohakuloa (open habitat) during June/July, 1992 Abbreviations are the same as in Figure 4.1a. 100
90 80
lU 70 (J) ~ c 60 lU ~ lU 50 a. cro 40 lU 1: ..... 30 ~ 20 10 0 0-2.0 2.1-4.0 4.1-6.0 6.1-10.0 10.1-14.0 14.1-18.0 18.1-24.0 length class (rnrn)
Figure 4.7. Size (body length) distribution (mean monthly percentage) of nocturnal aerial insects trapped at Ocean View (closed habitat), on the island of Hawai'i. n = 4 months. 100
90 eo 70 ....ro ....0 60 "- 0 ~ 50 ai u 40 L-eu a...... 30 N ..... 20 10 0
0-2.0 2.1-4.0 4.1-6.0 6.1-10.0 10.1-14.0 14.1-18.0 18.1-24.0
Length class (rnrn)
Figure 4.8. Size (body length) distribution of nocturnal aerial insects trapped at Red Cinder Road (closed habitat), on the island of Hawai'i. n = 1 month. 100
90
80
III 70 Ol ....ro c 60 Q) u L. II> 50 0. roC 40 III 1: 30
I-" 20 ~ 10 0
0-2.0 2.1-4.0 4.1-6.0 6.1-10.0 10.1-14.0 14.1-18.0 18.1-24.0
Length classes (rnrn)
Figure 4.9. Size (body length) distribution (mean monthly percentage) of nocturnal aerial insects trapped at Pohakuloa (open habitat), on the island of Hawai'i. n = 3 months. L. c. semotus
L. c. cinereus
E. fuscus
..... ~ L. nocttveaens
o 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Body length of Insect prey (mm)
Figure 5.1. Size range of insect prey taken by the Hawai'ian hoary bat, L. c. semotus, the North American hoary bat, L. c. cinereus, and putative competitors of the North American hoary bat, Lasionycteris noctivagans and Eptesicus fuscus. 120
_ 100~ '* ?fl. * -en J: 80 1 x -0 E tJ) 60 + + c L. c. clnereus :;: * + Competitors z Q)"' 40 en + x L. c. semotus-closed ~ II!- ~ "' 20 0 L. c. semotus-open , 0' I I I 0 20 40 60 80 100 Bats eating beetles (%)
Figure 5.2. Proportion of bats eating moths plotted against the proportion of bats eating beetles for the Hawai'ian hoary bat, L. c. semotus, the North American hoary bat, L. c. cinereus, and putative competitors of the North American hoary bat, Eptesicus fuscus, Myotis evotis, Myotis thysanodes. APPENDIX MUSEUM SKULL SPECIMEN NUMBERS AND STANDARD ERRORS OF SKULL MEASUREMENTS Table AI.
Museum skull specimens of the Hawai\ian hoary bat, Lasiurus cinereus semotus, and the North American hoary bat, Lasiurus cinereus cinereus, from which skull measurements were taken (Courtesy of the Museum of Southwestern Biology, New Mexico). Lasiurus cinereus semotus Lasiurus cinereus cinereus 43780 through 43789 10473 through 10477 43791 10483 through 10486 43792 through 43798 11747 and 11748 43800 11769 and 11770 43802 12802 through 12808
Table A2.
Standard errors for repeated measures (n = 10) taken from a single skull of the Hawai \ ian hoary bat, Lasiurus cinereus semotus. Measurement Standard error Dentary length 0.01 Length of masseter muscle scar 0.02 Height of coronoid process 0.01 Length of maxillary tooth row 0.02
125 LITERATURE CITED
Andrewartha, H. G. and Birch, L. C. 1955. Distribution and abundance of animals. Chicago Univ. Press.
Armstrong, R. W. ed. 1973. Atlas of Hawaii. Department of Geography, University of Hawaii. The University Press of Hawaii, Honolulu.
Baldwin, P. H. 1950. Occurrence and behavior of the Hawaiian bat. J. Mammal. 31(4):455-456.
Banko, W. E. and P. C. Banko. 1976. Role of food depletion by foreign organisms in historical decline of Hawai \ ian forest birds. In: Proceedings: First conference in natural sciences: Hawai \ i Volcanoes National Park. Coop. Natl. Park Res. Stud. Unit, Univ. Hawai \ i, Honolulu. C. W. Smith (ed.).
Barbour, R W. and W. H. Davis. 1969. Bats of America. University Press of Kentucky, Lexington.
Barclay, R M. 1985. Long-versus short range foraging strategies of hoary (Lasiurus cinereus) and silver-haired (Lasionycteris noctivagans) bats and the consequences for prey selection. Can. J. Zool. 63: 2507-2515.
------. 1986. The echolocation calls of hoary (Lasiurus cinereus) and silver haired (Lasionycteris noctivagans) bats as adaptations for long versus short-range foraging strategies and the consequences for prey selection. Can. J. Zool. 64:2700-2705.
------. 1988. Echolocation strategies of aerial insectivorous bats and their influence on prey selection. In: P. E. Nachtigall and P. W. B. Moore, (eds.). Animal Sonar. Plenum Press, New York.
Barclay, RM. and R. M. Brigham. 1991. Prey detection, dietary niche breadth, and body size in bats: why are aerial insectivorous bats so small? Am. Nat. 137(5): 693-703.
Belwood, J. and J. H. Fullard. 1984. Echolocation and foraging behavior in the Hawaiian hoary bat, Lasiurus cinereus. Can. J. Zool. 62: 2113-2120.
Black, H. L. 1972. Differential exploitation of moths by the bats Eptesicus fuscus and Lasiurus cinereus. J. Mamm. 53:599-601.
126 ------. 1974. A north temperate bat community: structure and prey populations. J. Mamm. 55:139-157.
Blair, W. F. 1974. Character displacement in frogs. Amer. ZooI.14: 1119-1125.
Blumenstock, D., S. Price. 1972. Climates of the states: Hawaii. In: Natural History of the Hawaiian islands. E. A. Kay (ed.) Univ. Hawaii Press, Honolulu.
Borror, D. J. 1981. An introduction to the study of insects. Saunders College Pub. Philadelphia.
Brown, J. H. 1971. Mammals on mountain tops: Nonequilibrium insular biogeography. Am. Nat., 105:467-478.
Brown, W. L. and E. O. Wilson. 1956. Character displacement. Syst. ZooI. 5:49-64.
Bryan, E. H., Jr. 1955. The Hawaiian bat. Elepaio 15:63-64.
Burkhardt, F. and S. Smith (eds.). 1991. The correspondence of Charles Darwin. Volume 7. Cambridge University Press, Cambridge.
Cameron, A. W. 1964. Competitive exclusion between the rodent genera Microtus and Clethrionomys. Evolution 18:630-634.
Carlquist, S. 1965. Island Life: A natural history of the islands of the world. Natural History Press, Garden City.
------. 1974. Island biology. Columbia University Press, New York.
Charnov, E. L. 1976. Optimal foraging: The Marginal Value Theorem. Theor. Pop. BioI. 9:129-136.
Cheverud, J. M. 1988. A comparison of genetic and phenotypic correlations. Evolution, 42:958-968.
Connell, J. H. 1980. Diversity and the evolution of competitors, or the ghost of competition past. Oikos 35: 131-138.
------. 1983. On the prevalence and relative importance of interspecific competition: Evidence from field experiments. Am. Nat. 122(5):661 696.
127 Cook, R. M. and B. J. Cockrell. 1978. Predator ingestion rate and its bearing on feeding time and the theory of optimal diets. J. Anim. Ecol. 47:529 547.
Crowell, K. L. 1962. Reduced interspecific competition among the birds of Bermuda. Ecology 43:75-87.
Darwin, C. 1950. On the origin of species by means of natural selection. Watts and Co., London. First reprint of the first edition published in 1859.
Davies, N. B. 1977. Prey selection and the search strategy of the Spotted Flycatcher (Muscicapa striata): a field study on optimal foraging. Anim. Behav. 25:1016-1033.
Davies, I. J. 1980. Relationships between dipteran emergence and phytoplankton production in the experimental lakes area, northwestern Ontario. Canadian Journal of Fisheries and Aquatic Sciences 37:523-533.
Den Boer, P. 1986. The present status of the competitive exclusion principle. Tree 1: 25-28.
Diamond, J. M. 1970. Ecological consequences of island colonization by southwest Pacific birds 1: Types of niche shifts. Proc. Nat. Acad. Sci. USA 67:529-536.
------. 1973. Distributional ecology of New Guinea birds. Science, 179: 759-769.
------. 1978. Niche shifts and the rediscovery of interspecific competition. Am. Sci. 66: 322- 31.
Dunham, A E., G. R. Smith, and J. N. Taylor. 1979. Evidence for ecogical character displacement in Western American catostomid fishes. Evolution 33 (3):877-896.
Duvall, F., R. Gassman-Duvall. 1991. No bats on Maui? Look again. Elepaio 51:1 2.
Eldredge, N. and S. J. Gould. 1972. Punctuated equilibria: an alternative to phyletic gradualism. In: T. J. M. Schopf (ed.), Models in Paleobiology.
Fagerstrom, J. A. 1978. Paleobiological application of character displacement and limiting similarity. Syst. Zool. 27(4):468-472.
128 Fairbairn, D. J. 1992. The origins of allometry: size and shape polymorphism in the common waterstrider, Gerris remigis Say (Heteroptera, Gerridae). Bio. J. Linn. Soc., 45:167-186.
Farney, J. and E. Fleharty. 1969. Aspect ratio, loading, wing span and membrane area of bats. J. Mamm. 50:362-367.
Fenchel, T. 1975. Character displacement and coexistence in mud snails (Hydrobiidae). Oecologia 20: 19-32.
Findley, J. S., E. H. Studier, and D. E. Wilson. 1972. Morphologic properties of bat wings. J. Mamm. 53:429-444.
Findley, J, S. and D. E. Wilson. 1982. Ecological significance of chiropteran morphology. In: Ecology of bats. T. H. Kunz (ed.). Plenum Press, New York.
Fitzpatrick, J. W. 1981. Search strategies in of Tyrant Flycatchers. Anim. Behav. 29:810-821.
Fjeldsa, J. 1982. Ecological character displacement and character release in grebes Podicipedidae. Ibis 125:463-481.
Foster, J. B. 1964. Evolution of mammals on islands. Nature 202:234-235.
Freeman, P. W. 1979. Specialized insectivory: Beetle-eating and moth-eating molossid bats. J. Mamm. 60(3):467-479.
------. 1981. Correspondence of food habits and morphology in insectivorous bats. J. Mamm. 62(1):166-173.
Fujioka, K. and S. M. Gon 111. 1988. Observations of the Hawaiian bat (Lasiurus cinereus semotusi in the districts of Ka 'u and South Kona, Island of Hawaii. J. Mamm. 69: 369-371.
Fullard, J. H. 1984. Acoustic relationships between tympanate moths and the Hawaiian hoary bat (Lasiurus cinereus semotus). J. Compo Physiol. 155: 795-801.
------. 1987. Sensory ecology and neuroethology of moths and bats: interactions in a global perspective. In: Recent advances in the study of bats. Fenton, M. B., P. Racey, J. M. V. Rayner (eds.). Cambridge University Press, Cambridge.
129 Futuyma, D. F. 1986. Evolutionary biology. Sinauer Associates, Inc., Sunderland.
Giller, P. S. 1980. The control of handling time and its effects on the foraging strategy of a heteropteran predator, Notonecta. J. Anim. Ecol. 49:699-712.
Gould, S. J. 1971. Geometric similarity in allometric growth: a contribution to the problem of scaling in the evolution of size. Am. Nat. 105:113-136.
------. 1975. On the scaling of tooth size in mammals. Amer. Zool, 15:351-362.
Gould, S. J. and R. C. l.ewontin. 1979. The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme. Proc. Roy. Soc. Series B 205:581-598.
Gould, S. J. and E. S. Vrba. 1982. Exaptation - a missing term in the science of form. Paleobiology, 8(1):4-15.
Grant, P. 1965. The adaptive significance of some size trends in island birds. Evolution 19:355-367.
------. 1966. The coexistence of two wren species of the genus Thyrothorus. Wilson Bulletin 78:266-278.
------. 1972. Convergent and divergent character displacement. BioI. J. Linn. Soc. 4: 39-68.
------. 1986. Ecology and evolution of Darwin I s finches. Princeton University Press, Princeton.
Grant, P. and D. Schluter. 1984. Interspecific competition inferred from patterns of guild structure. In: Ecological communities. Strong, D., D. Simberloff, L. G. Abele and A Thistle (eds.) Princeton Univ. Press.
Griffin, D. R. 1958. Listening in the dark. The acoustic orientation of bats and men. Yale University Press, New Haven.
------. 1971. The importance of atmospheric attenuation for the echolocation of bats (Chiroptera). Anim. Behav. 19:55-61.
Harvey, P. H. and M. D. Pagel. 1991. The comparative method in evolutionary biology. Oxford University Press, Oxford.
130 Hassell, M. P., J. H. Lawton, and J. R. Beddington. 1976. The components of arthropod predation I. The prey death-rate. J. Anim. EcoI. 45:135 164.
Hawaiian Heritage Program. 1992. Database of rare and endangered plants, animals, and natural communities of the Hawi \ ian islands. The Nature Conservancy, Honolulu, Hawaii. Unpublished.
Huey, R. B. and E. R. Pianka. 1974. Ecological character displacement in a lizard. Amer. ZooI., 14:1127-1136.
Husar, S. L. 1976. Behavioral character displacement: evidence of food partitioning in insectivorous bats. J. Mamm. 57:331-338.
Humphreys, W. J. 1940. Physics of the air. McGraw-Hill, Inc., New York.
James, H. F. and S. L. Olson. 1991. Descriptions of thirty-two new species of birds from the Hawai \ ian islands: Part II. Passeriformes. Ornith. Mon. 46:1-88.
Johnston, T. D. and G. Gottlieb. 1990. Neophenogenesis: A developmental theory of phenotypic evolution. J. Theor. BioI. 147:471-495.
Jones, C. 1965. Ecological distribution and activity periods of bats of the Mogollon Mountains area of New Mexico and adjacent Arizona. Tulane Studies in Zoology 12: 93-100.
Keast, A. 1967. Competitive interactions and the evolution of ecological niches as illustrated by the Australian honeyeater genus Melithreptus (Melipgagidae). Evolution 22: 762-784.
Kepler, C. B., J. M. Scott. 1990. Notes on distribution and behavior of the endangered Hawaiian hoary bat (Lasiurus cinereus semotus), 1964 1983. \ Elepaio 50 (7):1-6
Kick, S. A. 1982. Target-detection by the echolocating bat, Eptesicus fuscus. Journal of Comparative Physiology, 145:431-435.
Kleinbaum, D. G., K L. Kupper, and K E. Muller. 1988. Applied regression analysis and other multivariable methods. PWS-Kent Publishing Co., Boston.
Krebs, J. R. and N. B. Davies. 1978. Behavioural ecology: An evolutionary approach. Sinauer Associates, Sunderland.
131 Krebs, J. R., J. T. Erichsen, M. I. Webber, and E. L. Chamov. 1977. Optimal prey selection on the great tit (Parus major). Anim. Behav. 25:30-38.
Kunz, T. H. 1973. Resource utilization: temporal and spatial components of bat activity in central Iowa. J. Mamm. 54:14-32.
LaBarberra, M. 1989. Analyzing body size as a factor in ecology and evolution. Ann. Rev. Ecoi. Syst. 20:97-117.
Lack, D. 1947. Darwins finches. Cambridge Univ. Press.
------. 1976. Island biology illustrated by the land birds of Jamaica. University of California Press, Berkeley.
Lande, R. 1979. Quantitative genetic analysis of multivariate evolution, applied to brain:body size allometry. Evolution, 33(1):402-416.
Lawlor, T. E. 1982. The evolution of body size in mammals: Evidence from insular populations in Mexico. Am. Nat. 104:155-174.
Lister, A. M. 1989. Rapid dwarfing of red deer on Jersey in the last interglacial. Nature 342:539-542.
Lomolino, M. V. 1985. Body size of mammals on islands: The island rule re examined. Am. Nat. 125:310-316.
Lucas, J. R. and A. Grafen. 1985. Partial prey consumption by ambush predators. J. Theor. BioI. 113:455-473.
MacArthur, R. H., J. M. Diamond, and J. R. Karr. 1972. Density compensation in island faunas. Ecology, 53(2):330-342.
MacArthur, R. H. and E. O. Wilson. 1963. An equilibrium theory of insular zoogeography. Evolution, 17:373-387.
Marshall, L. G. and R. S. Corruccini. 1978. Variability, evolutionary rates, and allometry in dwarfing lineages. Paleobiology 4(2):101-119.
Maurer, B. A., J. H. Brown, and R. D. Rusler. 1992. The micro and macro in body size evolution. Evolution, 46(4):939-953.
Maynard Smith, J. and R. J. G. Savage. 1959. The mechanics of mammalian jaws. School Sci. Rev. 40:289-301.
132 Mayr, E. 1954. Change of gentic environment and evolution. In: Evolution as a process, pp.157-180. J. Huxley, A. C. Hardy, and E. B. Ford (eds.). Macmillan, New York.
McMahon, T. 1973. Size and shape in biology. Science 179:1201-1204.
McNab, B. K. 1971. On the ecological significance of Bergmann •s rule. Ecology 52 (5): 845-854.
McNair, J. N. 1983. A class of patch-use strategies. Amer. Zool. 23:303-313.
Mohl, B. 1988. Target detection by eeholocating bats. In: Animal sonar: Processes and Performance. NachtigaIl, P. E. and P. W. B. Moore (eds.). Plenum Press, New York.
Montgomery, S. L. 1983. Carnivorous caterpillars: The behavior, biogeography, and conservation of Eupithecia (Lepidoptera: Geometridae) in the Hawai \ ian islands. Geojournal 7.6:549-556.
Mountainspring, S. 1987. Ecology, behavior, and conservation of the Maui parrotbill. Condor, 89:24-39.
Mountainspring, S. and J. M. Scott. 1985. Interspecific competition among Hawai \ ian forest birds. Eeol. Monogr. 55:219-239.
Mueller-Dombois, D., K. W. Bridges and H. L. Carson. 1981. Structural variation in rainforest. In: Island ecosystems: Biological organization in selected Hawai \ ian communities. MueIIer-Dombois, D., K. W. Bridges, and H. L. Carson (eds.). Hutchinson Ross Pub. Co.
Norberg, U. M. 1981. Allometry of bat wings and legs and comparison with bird wings. Phil. Trans. R. Soc. Lond. B 292:33-398.
------. 1986. Evolutionary convergence in foraging niche and flight morphology in insectivorous aerial-hawking birds and bats. Om. Scand. 17:253-260.
------. 1990. Vertebrate flight: mechanics, physiology, morphology, ecology, and evolution. Springer-Verlag, Berlin.
Norberg, U. M. and M. B. Fenton. 1988. Carnivorous bats? Bio. J. Lin. Soc. 33:383-394.
133 Norberg, U. M. and J. M. V. Rayner. 1987. Ecological morphology and flight in bats (Mammalia; Chiroptera): Wing adaptations, flight performance, foraging strategy and echolocation. Phil. Trans. R. Soc. Lond. B 316:335-427.
Peters, R. H. 1983. The ecological implications of body size. Cambridge University Press, Cambridge.
Rabol, J. 1987. Coexistence and competition between overwintering willow warblers Phylloscopus trochilus and local warblers at Lake Naivasha, Kenya. Ornis Scand. 18: 101-121.
Riess, M. J. 1989. The allometry of growth and reproduction. Cambridge University Press, New York,
Riska, B., T. Prout, and M. Turelli. 1989. Laboratory estimates of heritability and genetic correlations in nature. Genetics, 123:865-871.
Roeder, K D. 1967. Nerve cells and insect behavior. Harvard University Press, Cambridge.
Ross, A. 1967. Ecological aspects of the food habits of insectivorous bats. Proc. Western Found. Vert. Zool. 1:204-263.
Sanborn, C. C. and J. A. Crespo. 1957. EI Murcielago Blanquizco (Lasiurus cinereus) y Sus Subespecies [The hoary bat (Lasiurus cinereus) and its subspecies]. Museo Argentino Ciencias Naturales "Bernardino RivadaviaII Bull. 4. 13pp. (In Spanish).
Schluter, D. 1986. Character displacement between distantly related taxa? Finches and bees in the Galapagos. Am. Nat. 127: 95-102.
Schluter, D., T. D. Price, and P. R. Grant. 1985. Ecological character displacement in Darwin's finches. Science 227:1056-1058.
Schmidt-Nielsen, K 1984. Scaling. Why is animal size so important? Cambridge University Press, New York.
Scholander, P. F. 1955. Evolution of climatic adaptation in homeotherms. Evolution 9:15-26.
Schoener, T. W. 1970. Size patterns in West Indian Anolis lizards. II. Correlations with the sizes of particular sympatric species-displacement and convergence. Amer. Nat. 104:155-175.
134 Sih, A. 1980. Optimal foraging:Partial consumption of prey. Am. Nat. 116(2):281 290.
Simon, C. 1987. Hawai-ian evolutionary biology: an introduction. Tree 2(7):175 178.
Simmons, J. A, W. A Lavender, B. A Lavender, J. E. Childs, K Hulebak, M. R. Rigden, J. Sherman, and B. Woolman, and M. J. 0'Farrell. 1978. Echolocation by free-tailed bats (Tadarida). J. Compo Phys. 125:291 299.
Smith, R. L. 1980. Ecology and field biology. Harper and Row, New York.
Sohmer, S. H. and R. Gustafson. 1987. Vegetation zones, pp.38-57 in Plants and Flowers of Hawai 'i. University of Hawai 'i Press, Honolulu.
Sondaar, P. Y. 1977. Insularity and its effect on mammalian evolution. In: Major patterns in vertebrate evolution. Hecht, M. K., P. C. Goody, and B. M. Hecht (eds.). Plenum Press. New York.
Stephens, D. W. and J. R. Krebs. 1986. Foraging theory. Princeton University Press, Princeton.
Tomich, P. Q. 1974. The Hawaiian hoary bat: daredevil of the volcanoes. Nation. Parks and Conserve Mag. 48:10-13.
------. 1986. Mammals in Hawaii. Bishop Museum Press, Honolulu.
Van Valen, L. 1973. Pattern and the balance of nature. Evol. Theory, 1:31-49.
Vaughan, T. A. 1966. Morphology and flight characteristics of molossid bats. J. Mamm. 47(2):249-260.
Walker, T. J. 1974. Character displacement and acoustic insects. Amer. Zool. 14: 1137-1150.
Warner, R. M. 1985. Interspecific and temporal dietary variation in an Arizona bat community. J. Mamm. 66(1): 45-51.
Werner, E. E. and D. J. Hall. 1974. Optimal foraging and the size selection of prey by the bluegill sunfish Lepomis macrochirus. Ecology 55:1042 1052.
135 Whitaker, J. O. 1972. Food habits of bats from Indiana. Can. J. Zoo!' 50: 877-883.
Whitaker, J. O. and P. Q. Tomich. 1983. Food habits of the hoary bat, Lasiurus cinereus from Hawaii. J. Mamm. 64:151-152.
Williams, D. F. and J. S. Findley. 1979. Sexual size dimorphism in Verspertilionid bats. Am. MidI. Nat. 102(1):113-126.
Williams, E. E. 1969. The ecology of colonization as seen in the Zoogeography of anoline lizards on small islands. Quart.Rev. BioI. 44(4):345-389.
Williams, G. C. 1966. Adaptation and natural selection. Princeton University Press, Princeton, New Jersey.
Zimmerman, E. C. 1970. Adaptive radiation in Hawai \ i with special reference to insects. Biotropica 2(1):32-38.
Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall, Inc., New Jersey.
136