Aspects of the reproductive biology of two carpenter bees (genus Xylocopa) in southern Arizona
Item Type text; Thesis-Reproduction (electronic)
Authors Minckley, Robert Lynn, 1957-
Publisher The University of Arizona.
Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
Download date 29/09/2021 14:22:02
Link to Item http://hdl.handle.net/10150/276573 INFORMATION TO USERS
This reproduction was made from a copy of a document sent to us for microfilming. While the most advanced technology has been used to photograph and reproduce this document, the quality of the reproduction is heavily dependent upon the quality of the material submitted.
The following explanation of techniques is provided to help clarify markings or notations which may appear on this reproduction.
1.The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure complete continuity.
2. When an image on the film is obliterated with a round black mark, it is an indication of either blurred copy because of movement during exposure, duplicate copy, or copyrighted materials that should not have been filmed. For blurred pages, a good image of the page can be found in the adjacent frame. If copyrighted materials were deleted, a target note will appear listing the pages in the adjacent frame.
3. When a map, drawing or chart, etc., is part of the material being photographed, a definite method of "sectioning" the material has been followed. It is customary to begin filming at the upper left hand corner of a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the first row and continuing on until complete.
4. For illustrations that cannot be satisfactorily reproduced by xerographic means, photographic prints can be purchased at additional cost and inserted into your xerographic copy. These prints are available upon request from the Dissertations Customer Services Department.
5. Some pages in any document may have indistinct print. In all cases the best available copy has been filmed.
University MicKxilms International 300 N. Zeeb Road Ann Arbor, Ml 48106
Order Number 1332421
Aspects of the reproductive biology of two carpenter bees (genus Xylocopa) in southern Arizona
Minckley, Robert Lynn, M.S.
The University of Arizona, 1987
UMI 300 N. Zeeb Rd. Ann Arbor, MI 48106
PLEASE NOTE:
In all cases this material has been filmed in the best possible way from the available copy. Problems encountered with this document have been identified here with a check mark V .
1. Glossy photographs or pages
2. Colored illustrations, paper or print
3. Photographs with dark background
4. Illustrations are poor copy
5. Pages with black marks, not original copy
6. Print shows through as there is text on both sides of page
7. Indistinct, broken or small print on several pages.
8. Print exceeds margin requirements
9. Tightly bound copy with print lost in spine
10. Computer printout pages with indistinct print
11. Page(s) lacking when material received, and not available from school or author.
12. Page(s) seem to be missing in numbering only as text follows.
13. Two pages numbered . Text follows.
14. Curling and wrinkled pages
15. Dissertation contains pages with print at a slant, filmed as received
16. Other
ASPECTS OF THE REPRODUCTIVE BIOLOGY OF TWO
CARPENTER BEES (GENUS XYLOCOPA) IN
SOUTHERN ARIZONA
by
Robert Lynn Minckley
A Thesis Submitted to the Faculty of the
DEPARTMENT OF ENTOMOLOGY
In Partial Fulfillment of the Requirements for the Degree of
MASTER OF SCIENCES
In the Graduate College
THE UNIVERSITY OF ARIZONA
19 8 7 STATEMENT BY AUTHOR
This thesis has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.
Brief quotations from this thesis are allowable without special permission, provided that accurate acknowledgement of source is made. Request for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgement the proposed use of the material is in the interests os scholarship. In all other instances, however, permission must be obtained from the author.
SIGNED:
APPROVAL BY THESIS DIRECTOR
has been approved on the date shown below:
R. L. SMITH iociate Professor of Entomology ACKNOWLEDGEMENTS
I wish to express my gratitude to the members of my advisory
committee, Or. Villiam Nutting for providing many helpful suggestions
and criticisms of my research, Dr. Robert Smith who gave me the
opportunity to pursue my ideas freely and in particular Dr. Stephen
Buchmann for his constant patience, insight and support.
Numerous persons contributed to completion of this study but three were abundantly helpful throughout and are due specific acknowledgement,
Lawrence Puis and Micheal Calhoun who helped in field work and Donald
Gilmour who did the illustrations.
Finally, I thank my wife, Nancy, who contributed enormously to every facet of this project.
ill TABLE OF CONTENTS
Page
LIST OF TABLES vii
LIST OF ILLUSTRATIONS ix
ABSTRACT xi
CHAPTER I. NESTING ECOLOGY OF XYLOCOPA CALIFORNICA ARIZONENSIS 1
Introduction 1
General Biology of Xylocopa callfornica arlzonensis 2
Nest Site Selection 4
Materials and Methods 5
Description of Study Site 5
A. palmeri Life History 6
A. palmeri Phenology 6
Nesting 8
Nest Woods 9
Results 10
A. palmeri Phenology 10
Stalk Attrition 16
Nest Site Selection 19
Nest Microsites 21
Substrate Properties 23
iv V
Substrate Availability 28
Substrate Usage Patterns 30
Discussion 34
Nest Site Selection and the Thermal Environment 37
Limiting Factors on Nesting Success 41
Speciation in the Xylocopini 42
Summary 47
CHAPTER II. LIFE HISTORY, NESTING BIOLOGY AND MALE MATING
BEHAVIOR OF THE LARGE CARPENTER BEE, XYLOCOPA (NEOXYLOCOPA)
VARIPUNCTA (HYMENOPTERA: ANTHOPHORIDAE) 48
Introduction 48
Description of the study site 51
Materials and Methods 52
General activity patterns and life history 52
Male mating behavior 52
Results 53
Nest substrate usage and nest architecture 53
Life history 53
Behavior at territories 57
Display site selection 64
Spatial patterns 65
Discussion 69
Life history, nesting biology and nest architecture 69 vi
Male mating behavior 71
Site usage patterns and territorial placement tactics .... 79
LITERATURE CITED 85 LIST OF TABLES
Table Page
1 List of nesting substrata known for X. c. arizonensis 3
2 Analysis of stalk use by X. c. arizonensis during 1985
and 1986 in the Rincon Mts.; Spring includes Mar.,
Apr., and May and summer includes June, July, and
August 23
3 Breakdown of stalk conditions (explained in text) and
carpenter bee usage in April, 1986 28
4 Thermal conductivity and density measurements for 3
wood types 30
5 Factors correlated with nest abandonment in spring,
1986 34
6 Length of nest occupation in stalks kept in the
laboratory 34
7 Nesting substrates used by X. varipuncta 55
8 Types of male-male aggressive encounters and outcomes.
See text for explanation of encounter types 64
vii LIST OF ILLUSTRATIONS
Figure Page
1 Nesting behavior of Xylocopa californica arizonensls
at Tucson, Az, 1986. Dashed lines indicate that
events were not observed at the study site but
probably occurred, solid lines are recorded
observations 8
2 Laboratory setup for determination of thermal
conductivity measurements 12
3 Number of Agave palmer! stalks from January 1985 to
September 1986 15
4 Standing floral scapes from January 1985 to September
1986 separated according to age class 16
5 Graph of stalk attrition through time for the 1984
age class 18
6 Probability a stalk will fall before the next census
based on the 1984 age class (n = 66) plotted against
the actual numbers which fell each month. Individual
variation occurs within the population and for
purposes of this graph are all assumed to flower on
August 1 19
viii ix
7 A breakdown of the standing scapes in July 1986
based on condition and age. Condition 1 = green stalk
still elongating; condition 2 = green stalk with
flower buds; condition 3 = green stalk with flowers;
condition 4 = green stalk with fruit; condition 5 =
green stalk with dried seed pods; condition 6 =
dried stalk with dried seed pods 21
8 Proposed scenario for the combined effect of
decomposer and herbivore/ xylophile life histories
on stalk quality plotted with the attrition pattern
of the 1984 Agave palmeri age class 25
9 Temperature profile on 6 June, 1985 of Xylocopa
californica arizonensis nest in a larval cell and in the
nest tunnel 26
10 Map of study area showing lek and nest sites. Refer
to figure 5 for letter designations 56
11 Life history of Xylocopa varipuncta 60
12 Temperature data for month preceding start of
seasonal activity for 1984, 1985 and 1986 61
13 Rate of female return flights and percent of trips
with pollen 62
14 1984 site and station use showing flowering phenology
of focal plants 70 X
15 Frequency of male display times for 53 males from
1984 and 1985 mating seasons 72
16 Frequency of number of sites used per day 74
17 Number of days sites were used in 1984 mating season 75
18 Frequency preferred site was used compared with
frequency of the number of sites used per day with
the chinaberry site excluded 76 ABSTRACT
Two species of large carpenter bees (genus Xylocopa) were studied in southern Arizona. Nesting preference of X. californica arlzonensls in floral scapes of Agave palmerl was found to depend on floral scape age and status of the thermal microenvironment. Evidence suggests that females assess scape age with extreme precision and that this ability enables them to avoid those scapes degraded by other biota. Females also preferentially used Agave scapes that were shaded from the afternoon sunlight. Male mating behavior and lek site selection of X. varipuncta was also investigated near a large nesting site in a topographically flat area. Behaviors analyzed indicate that males tend to display singly and that they disperse pheromonal signals. Lek site selection was in accord with that found at three other study sites and further demonstrated that males did not clump their territories near nest sites.
Distribution of male territories therefore does not appear to correspond to female distributions in the environment.
xi CHAPTER 1
NESTING ECOLOGY OF XYLOCOPA CALIFORNICA ARIZONENSIS
Introduction
Carpenter bees of the subfamily Xylocopinae construct their nests
almost exclusively in wood. Three tribes make up the subfamily: the
allodapine bees (Allodapini), the twig nesting carpenter bees
(Ceratinini) and the large carpenter bees (Xylocopini)(Daly et al,
1987). The former group is composed of three genera of which one,
Xylocopa, is by far the most speciose (Hurd and Moure, 1963). Xylocopa
is worldwide in distribution, with most species in the tropics (Hurd and
Moure, 1963; Michener, 1979). Nesting architecture and biology of many
xylocopid species have been described because, once located, nests can
be easily examined and the bees' large size makes them conspicuous and
easy to observe. Behaviors at nest sites are well known from X-ray
techniques (Gerling, Hurd and Hefetz, 1981, 1983) and external
observations (Anzenberger, 1979; Janzen, 1966). The wood types used for
nesting have been compiled for the New World species by Hurd (1978a) who
points out that nesting substrates used by one species are not specified
by plant type but seemingly more to the "quality" of the wood (i.e. low density, punky, etc). He further suggested that speciation and divergence are the result of differences in nest substrate preferences
(Hurd, 1958; Hurd and Moure, 1963, Hurd, 1978a). Adaptations to nesting
1 2
in wood thus provide a major theme to the investigators concentrating on
this genus. However, to date no study on nesting ecology has been
conducted for any species. Xylocopas spp. are widely distributed and
therefore subject to a variety of environmental regimes. This diversity
may therefore be important to understanding the variety of nesting substrates chosen and the number of species known.
The purpose of this paper is to delimit the environmental
features important in nest site selection of Xylocopa californica arizonensis and to discuss whether nest site selection may have served
to allow X. c. arizonensis to become genetically isolated from other subspecies.
General Biology of X. c. arizonensis
X. c. arizonensis ranges throughout the Sonoran and Chihuahuan deserts of southwestern North America from elevations below sea level in
Death Valley, California (Hurd, 1978) to 1500 m in the Santa Catalina
Mts. near Tucson (pers. obs.). Nests are typically single, non-branched tunnels but are found with multiple branches when material thickness permits (unpubl. data).
A large number of plant types are used for nesting that can be grouped into two large categories based on habitat requirements, either mesic or xeric-adapted. Mesophytic plants are locally abundant in the Sonoran Desert along waterways with permanent water. Xeric-adapted plants used for nesting all produce a woody scape bearing the reproductive structures of the plant. These occur patchily within their Table 1. List of nesting substrata known for X. c. arizonensis
Substrate Reference
Fremonts' cottonwood Hurd, 1978; this study Populus fremontii
California fan palm Hurd, 1978; this study Washingtonia filifera
Willow Hurd, 1978 Sallx spp.
Elephant tree Hurd, 1978 Bursera microphylla
Agave spp. Hurd, 1978; this study
Bear grass Hurd, 1978 Nolina
Sotol Hurd, 1978; this study Dasylirion wheeleri
Soaptree Yucca Fowler, 1985; Hurd, 1956, 1978; Yucca elata this study
Redwood fence pickets Hurd, 1978; this study
Oleander this study Nerlum oleander
Saguaro rib this study Cereus carnegia 4
ranges but are more widely occurring than the mesic-adapted plant group.
Table 1 lists all nesting substrata known to be used by this subspecies.
Although other wood types may be available, X. c. arizonensis prefers
softer, structurally sound woods. No nests have been reported from the
widespread, but denser, wood of mesquite (Prosopis spp.) and palo verde
(Cercidium spp.) which occur throughout most of the bees' range.
Material must also be a minimum of 16 mm. in diameter to accomodate
these nests. The life history of X. c. arizonensis is similar to that
described for X. sulcatlpes in Israel (Gerling, Hurd and Hefetz, 1983;
shown diagrammatically in Fig. 1). Overwintering groups composed of
both sexes become active near the beginning of March in Arizona.
Weather at this time of year is transitional from the cold temperatures
of winter and can revert quickly, supressing any carpenter bee activity
for several days. Adult groups stay together until mid-March at which
time some females disperse to new nest sites and singly begin
construction of new tunnels. Males establish mating territories and
continue to sleep in the old nests after the ;:emales leave. Provisioning
and egg-laying begin in late March and development of the young lasts
around 45 days (Minckley, unpubl.). From 4 to 14 cells were found in
early season nests which appeared to be the product of a single female
(Minckley and Buchmann, unpubl.) The first generation of the year
emerges from their natal cells in early June and begin flying 10-20 days
later. Provisioning for the second generation occurs in some nests
during late July although there were nests which were not reprovisioned.
It is not clear whether females are capable of rearing a second brood or
whether females which deferred earlier reproduction are responsible for 5 the observed second brood. Although activity extends into November, no evidence indicates that there are more than two generations per year as has been reported for X. sulcatipes in Israel (Gerling, Hurd and Hefetz,
1983).
Carpenter bees have low fecundity and a protracted prereproductive period of up to 9 months (Gerling, Hurd and Hefetz,
1983; Skaife, 1952). Eggs are among the largest known for any insect (up to 14mm.)(Iwata, 1964). Female parental investment also includes guarding young and trophallactic feeding as adults (Gerling, Hurd and
Hefetz, 1983; Michener, 1972). All studies of male mating behavior in
Xylocopa involve the use and defense of territories (Marshall and
Alcock, 1981) and activity may last from a month to nearly the entire activity season. X. c. arizonensis males are active for 9 of the 10 month activity season in southern Arizona (Fig. 1).
Nest Site Selection
Relationships between climatic factors and nest site selection have been addressed for a variety of bee taxa. Clearly, high ambient temperature is the single factor correlating best with nest site placement. Barrows (1980) reported that carpenter bee nests of X. v. virginica in the Washington, D.C. area were positioned south and west significantly more often than in any other direction. In Brazil,
Michener et. al. (1958) noted a predominance of ground dwelling bees' nests facing north (toward the midday sun) and they theorized that the bees benefited by earlier activity made possible from the higher soil temperatures. Vluegal (1947), in Holland, showed that individuals of 6
Andrena vaga which nested in dikes facing southeast became active
earlier than those from nests facing west. In desert regions with higher
ambient temperatures nests seem to be positioned to avoid direct
sunlight. With an increase in latitude or elevation the combined effects
of solar radiation and ambient temperature decline linearly and in these
localities low temperature is likely to be more deleterious or lethal
(Seely and Visscher, 1985). Rau (1933) monitoring a population of X. v.
virginica reported 84£ mortality during one unusually cold winter.
The high blackbody temperatures consistently attained in deserts
are directly related to higher ambient temperatures. Because radiant
energy is fixed at any given time of the day the final temperature a
material attains varies according to the varation in ambient
temperature. This explains the similar curves for ambient and stalk
temperatures seen in Fig. 9. High ambient temperatures are necessary for
a nest to exceed physiologically lethal temperatures in animals (Pepper
and Hastings, 1952). In hot climates the heat load can regularly exceed
physiological thermal maxima. Nest locations appear to be chosen to
decrease deleterious effects of the radiant environment.
Materials and Methods
Description of Study Site
Field investigations were conducted from July 1984 through
September 1986 on a 14.4 hectare plot at Rincon Creek Ranch located on
the south-facing slopes of the Rincon Mountains, Pima Co., AZ. The site
is at an elevation of ca. 1070 m. and has been undisturbed by cattle at 7
least since 1979. Dominant perennial vegetation include palo verde
(Cercidium microphyllum (Torr.)), catclaw acacia (Acacia greggii Gray),
saghuaro (Cereus gigantea Engelm.), mesquite (Prosopis velutina
(Woot.)), triangle leaf bursage (Ambrosia deltoidea (Torr)) and Palmer's
agave (Agave palmerl).
A. palmeri Life History
At this study site the dried floral scapes of A. palmeri provide
the exclusive nesting substratum used by X. c. arizonensis. A. palmeri
is a desert plant which grows as a rosette and reproduces once
(semelparous) after reaching 8 to 25 years of age (Gentry, 1982). During
vegetative growth the leaves of the plant grow to a maximum of 1.5 m. as
they harness the energy needed for reproduction. Energy is stored
primarily in the bole of the plant, as starch, which at reproduction is
composed of over 50% carbohydrates (Howell and Roth, 1981). Blooming
occurs in the summer months and the appearance of flowers is preceded by
the growth of a stalk or scape up to 6 meters high (Gentry, 1982).
Carbohydrates are expended with the growth of the reproductive
structures and the plant withers. Once dried, stalks deteriorate rapidly
and their color darkens which allows a simple determination of age since
reproduction.
A. palmeri Phenology
All stalks were individually numbered so that their condition
could be monitored through time. Six categories of stalk condition were
designated 1) actively growing but lacking flower buds; 2) stalk growth
terminated, flower buds present; 3) stalk green, flowers present; &EAS.ON ACTIVITY
MALE TEf^PilTOMAL-ITY
MOW NefrT COKJ&TMJOTION
PP(PI¥I^I@NIN& Fp.ce&/f=p>As»&
f=eMAU6 UfcU^PATION
ANThhAX
2 3 4 e &> -7 a •? io ii MONTHS
Figure 1. Nesting behavior of Xylocopa calif ornica arizonensis at Tucson, AZ, 1986. Dashed lines indicate that events were not observed at the study site but probably occurred, solid lines are recorded observations. 9
4) stalk green, fruit present; 5) stalk dried; seeds present in pods; and 6) stalk dried, no seeds in pods. In summer, these categories often overlapped on one individual. In these cases, the latest stage of development was recorded. Cover or shading from adjacent vegetation was estimated subjectively on a scale of 0 to 3 (0 being completely unshaded and 3 denoting dense shading).
Nesting
Bee nest data included height of nest, compass orientation of the entrance, stalk diameter at entrance, presence of debris from inside the nest (wood shavings, pupal feces, exuviae) and whether the nest appeared active or not. Nest viability was often difficult to determine accurately due to decreased activity of the bees (particularly in the winter months) or recent predator damage. In these cases observations in later months ascertained the status.
For comparison, nests and unused flower scapes were removed and placed in shade with no ground contact. Individual nests were then monitored and data on life history recorded, particularly that pertaining to nesting activity and brood rearing. Nesting behaviors were classified as follows: 1) construction of new nests; 2) nest elongstion and subsequent cell provisioning indicated by the production of fresh wood shavings or frass below nest entrances; and 3) emergence of teneral adults indicated by the presence of pollen- soiled wood shavings, cell partition material, fecal pellets, exuviae and dead larval bees below nests. 10
Nest Woods
To determine whether differences in wood structure could affect
nesting success, thermal conductivity and wood density of A. palmeri,
Yucca elata Engelm. and Dasylirlon wheeleri Wats, were tested. These
three wood types are all commonly used for nesting by X. c arizonensis
in Arizona. Fig. 2 illustrates the apparatus used to test heat
conductivity of wood in the laboratory. A 9.3 liter tank with a lid was
filled with ice water and served as the "cold sink". A motor driven
stirrer set at 3.5 revolutions/ sec. maintained an even temperature
throughout the tank. A flat 5.1 cm. high x 5.1 cm. width section of the
wood was mounted flush to the side of the tank. A 5.1 x 5.1 cm. heat
flux meter was then attached to the outer side of the test sample. Two
0.64 mm copper - constantin thermocouples (TypelT-lE) connected to a
digital thermometer (Bailey Instrument, Model BAT-8) were inserted, one
between the heat flux meter and the sample, another between the tank and
the sample. The temperature was recorded after remaining stable for 15
minutes.
Coefficient of heat conductivity was calculated from the basic
heat transfer formula (Sears and Zemansky, 1960)
1 (1) Q = k A (tx - t2) T 1'
where Q represents energy; k is thermal conductivity of the material
being tested (here expressed as watts x °C x meter.j); A is area in 9 meters ; tj - t2 is the change in degrees measured as centigrade inside and outside, respectively and 1 the sample thickness in meters. With
the apparatus used in this experiment all values were known except for k. The equation when solved for k becomes 11
1 1 1 (2) k = Q 1 T' A" (t1-t2)*
The heat flux meter simplified the equation by giving a value Z which encompassed the variables Q T"* A'* and made the heat conductivity a three parameter equation
(3) k - Z 1 The inverse of k calculates resistance to thermal conductance or insulation of the material tested. Densities (weight x volume"^) of the three wood types were also calculated with weights obtained on a Mettler analytical balance. Field studies of temperature regimes were taken using an Omnidata Data Logger or the Sensortek Digital Thermometer, BAT-8. Results A. palmeri Phenology At the Rincon Creek Ranch study site, A. palmeri was associated nonrandomly with perennial plant species. Of the 72 post-reproductive A. palmeri in the 1985 age class, 69.4% were within the canopy of another plant (predominantly Cercidium microphyllum) and 30.6% were established in the open. The relationship is significant (X^=10.8; p>.01, df=l) and indicates that the nurse tree effect increases successful establishment of A. palmeri. This secondarily influences the nesting success of X- c. arizonensis (see below). Initiation of stalk growth in A. palmeri was from May to July. In mid-June flowering began and continued until September. Development of fruits continued from late June until late September. Typically once the green fruits appeared, stalk discoloration associated with drying was apparent. Drying is slow during the winter 12 DIGITAL THERMOMETER VOLT METER MOTOR DRIVEN STIRRER TEMPERATURE PROBES WOOD HEAT FLUX METER ICE WATER WATER TANK Figure 2. Operation setup for the determination of thermal conductivity of nesting substrata. 13 but by March most fruits had dehisced and flower scapes were brown- colored and dry, but still structurally intact. Figure 3 shows the total number of standing A. palmeri stalks present from 1984 to 1986. Maximum numbers of stalks were present during the flowering period in July and minimum numbers in June just prior to flowering. The yearly turnover of standing stalks was high. From July 1985 to June 1986, 43.4% of 133 total stalks fell. In general, low rates of stalk attrition were associated with autumn and winter when ambient temperatures decreased. Higher attrition rates occurred during spring and summer when temperatures were higher. Figure 4 shows the standing crop of flower scapes when separated according to age class. Stalks of various ages are present in all samples; however, the most recent age class - 0 to 1 year old - is numerically dominant. Second year stalks account for the majority of remaining stalks. Stalks older than 2 years were few in number, representing 10% or less of standing stalks. Zero to 1 and 1 to 2 year old stalk classes differ in their attrition curves (Fig. 5). One- to- two year old stalks fall more frequently than 0 to 1 year old stalks. Thus, the dried stalks present in a single sample will usually represent 3 age classes (Fig. 4: Fig. 7). All age classes follow a similar pattern of attrition. Fig. 5 tracks the 1984 age class through 1986, which for this discussion is considered to have flowered entirely in July. No stalks were observed to fall until they were more than 4 months old. The first major decline began at 12 months (i.e. August of the following year). At 12 months 90% 14 of the population remained standing yet in the following three months (age • 15 months) 50X of the population was gone and at 24 months fully 852 had fallen. Of 76 stalks in the 1985 cohort, 13 (17%) had fallen the following May, 1986 and by September, 1986, 27 (46%) had fallen (Fig. 4). In Fig. 6 the probability that a stalk will fall through time based on the 1984 age class is plotted. Since attrition curves for all age classes are similar the 1984 data set should be representative. Stalk falling rates were low in the winter after flowering and increase dramatically the following summer. High attrition occurred through the subsequent winter and began to decrease the following spring (Fig. 4). This decrease in attrition probability is an artifact created by stalks being supported by nearby trees. If all were unsupported, it appears that all members of one age class would fall in 17 to 22 months after the flowering event. Throughout its life cycle, A. palmeri offers a number of resources to invertebrate and vertebrate organisms (Howell and Roth, 1981; Schaffer and Schaffer, 1979; Waring and Smith, 1986). Associated with flowering are new plant structures and chemical changes of the leaf and bole. An example of the resource diversity available at one time is illustrated by the data for July 1986 (Fig. 7) a period when the greatest number of structures and ages are present. A small fraction of stalks are elongating (condition 1), 10.9% have flower buds (condition 2), relatively few are only flowering (condition 3) (although this is a deceiving product of the procedural design: simultaneously on one stalk 15 ICeO MONTHS' Figure 3. Number of Agave palmeri stalks from January 1985 to September 1986. 16 •zoo l&O U ICoO J j4 IAO 1L li0 0 100 £ 111 so £ ^ &>Q z 40 20 O I 2.34-5<£>78^IOII 12. I 2.^-4-&6»7&? I9e>3 | |90<£> MONTHS Figure 4. Standing floral scapes from January 1985 to September 1986 separated according to age class. 17 there can be bracts in flower, bud, and fruit) and the stalks with dehisced seed capsules (condition 5) were slow senescing members of the 1985 age class as were the dried stalks with no seeds (condition 6). The few stalks which were two or more years old had lost the vegetative material at the stalk base and were supported by adjacent trees. Associated communities can thus be segregated temporally, by being specific to one stage in plant development, or spatially, to one plant structure (Waring and Smith, 1986). However, since after flowering spatial and temporal features co-vary extensively the effect is best considered together. Stalk Attrition From the pattern shown in Fig. 5 sunlight, moisture and biotic decomposition may all contribute to stalk deterioration. Opened stalks revealed that decomposition was most prevalent within the soft interior pith and was caused by insect borings, indicating that biotic action and not sunlight or moisture were primarily responsible for structural failure. Two groups best characterized by trophic level are commonly observed in fallen stalks, herbivores/xylophiles and decomposers. Stalk structural quality is undoubtedly affected by both groups although the relative magnitude is unknown. Although life histories of decomposers (Mackay et al, 1986) and the herbivores /xylophiles (e.g. Powell and Mackie, 1964; Waring and Smith, 1986) which utilize floral scapes of Agave are generally unknown comparison suggests that floral scape falling results from an additive effect of the two groups and their activity rate varies seasonally. 18 z IOO Q z "9d , 7C> J 4 6,0 b SO 1L 0 40 1- 2 SO 111 0 zo £ it I 5! IO % ATTMTIOU 7 » <9 IO II 12. I 2.-54-S Figure 5. Graph of stalk attrition through time for the 1984 age class. 19 .24- CoCo ^ V—PfSOO/MblL-ITY ^ ^ Ol" 1»Al_I_IN<3r 60 54- A J 4-6 £ A2. ~Z 111 / /— NUMBCPi OF*- T=A«L_l_eSI STALKS 5 9<9 j 2.4- € ssi I. Ift 3 2 12. © 12. \Co ZO 22. Z.fc MONTHS AFTEPi »LOOMIN& Figure 6. Probability a stalk would fall before the next census based on the 1984 age class (n - 66) plotted against the actual numbers which fell each month. Individual variation occurs within the population and for purposes of this graph are all assumed to flower on August 1. 20 Initially, decomposers are few in number and herbivores are a small size and therefore have little weakening effect. Because colonization occurs in mid to late summer within 1 or 2 months, burrowing and decomposition activity is curtailed by low winter temperatures. This seasonal effect on the biota effectively diminishes the rate of winter stalk attrition. It is the following summer when the stalk fauna has a complete season ahead, that the combination of burrowing by herbivore-xylophiles and deterioration by decomposers may best explain the pattern of stalk attrition. This scenario is illustrated in Fig. 8. Nest Site Selection Usage patterns indicate that carpenter bees prefer to nest in the most structurally sound substrates available. Table 2 gives the ages of A. palmeri stalks used for nesting by X. c. arizonensis females used for nesting in 1985 and 1986. In 1986, 23 of 26 new nests were constructed in stalks of the most recent age class. Of the 21 nests constructed in April, all were in stalks that had developed the preceding year. Four nests were excavated in the latter half of the summer. Two were constructed in stalks that had bloomed earlier that summer and two in 1- year old stalks. 1985 data were not used in this analysis since nests in 2-year old stalks may have been constructed in 1984. However, here again most nests (26 of 32) were in the 1- year- old stalks. No observations of females initiating nests in stalks greater than 1- year- old were made during either 1985 or 1986. Two nesting periods in 1986 produced data suitable to test female selectivity based on substrate quality. In the spring when the first 21 Figure 7. A breakdown of the standing scapes in July 1986 based on condition and age. Condition 1 = green stalk still elongating; condition 2 = green stalk with flower buds; condition 3 = green stalk with flowers; condition 4 » green stalk with fruit; condition 5 » green stalk with dried seed pods; condition 6 =* dried stalk and dried seed pods. 22 nesting activity commences, green stalks are not present and all standing stalks are dried from the previous summer. In the late nesting period stalks from the current years' bloom are present and, though plants have begun to senesce, they are still green indicative of some continued photosynthetic activity. Different nesting substrates were chosen in the two nesting periods. In the spring, females chose 1-year old stalks only; however, in the late summer (August and September) nests were built in 2-month-old stalks (green) and 1-year old stalks. Unfortunately, so few nests were built during the second nesting period that it is difficult to draw conclusions. However, the use of the green stalk indicates that selectivity toward fresh material may have been exercised. Nest Microsites Flower scapes are subject to a wide range of temperatures in the summer months depending largely on their exposure to sunlight. Fig. 9 graphs the temperature change in one X. c. arizonensis nest on 6 June, 1985. Data were collected in Tucson, AZ. (elevation 900 m) in an exposed vertical stalk. Exposure to sunlight caused rapid changes in temperature and on this day nest temperatures exceeded 40°C for extended periods and peaked at 45°C. Shaded stalks are numerous because of the differential success of seedlings growing under perennial vegetation. Nurse tree effect therefore influences the substrate availability to carpenter bees. Of the 76 plants present in June of 1985, 85% grew close to a perennial plant, typically a palo verde (Cercidium microphyllum). As is apparent from Table 2 successful nests do also occur in unshaded A. 23 Table 2. Analysis of stalk use by X. c. arizonensis for 1985 and 1986 in the Rincon Mts.; spring includes Mar., Apr., and May and summer includes June, July and August. 1985 1986 1986 1986 SPRING SPRING SUMMER total # active holes 32 25 13 29 # of stalks with 2 holes 1 3 1 3 total no. stalks with nests 31 22 12 26 stalk cohorts 0 year age group 0 0 2 2 1 year age group 25 21 10 25 2 year age group 6 1 1 1 # of stalks abandoned before 30 16 ? ? next season stalks used new from previous 7 21 4 season 24 palmer! stalks. These nests are consistently located behind the dried leaves of A. palmeri and are far fewer in number than nests in shaded stalks and perhaps are discriminated against because the nest lengths are limited due to the thermal load of regions which extend beyond the protection of the leaves. Nests were constructed preferentially in stalks shaded from the afternoon sun (Table 3). I grouped the stalks of the 1985 age class in one category according to southerly or westerly exposure and in another category if shaded in these directions. This was intended to determine whether Xylocopa discriminate between stalks exposed and protected in these directions. Statistically, this revealed 9 of 23 (39.1%) protected stalks were used as nesting sites as compared to 5 of 38 (13.2%) exposed stalks. Bees, therefore, do choose shaded stalks signifigantly more often (Z = 20.97; p > 0.005). Shading is clearly important to nesting females. Even in exposed stalks, nests were located so they were always shaded by the dried leaves of the plant. Substrate Properties To test whether the thermal conductivity and/or structural composition of scape pith used by X. c. arizonensis might differ enough to influence nest site selection, measurements on the three most common nest substrata were made. It could be that nests are excavated on the least exposed side of a large stalk to take advantage of the insulative properties of the wood on the exposed side. Thermal conductance and densities from these wood types are shown in Table 4 along with pine lumber as a reference. D. wheeleri and Y. elata are long-lived perennial 25 ifr COMBINES CFfECT 2 HEPS&IVOPiE/X^L-OPHIl-eS) 5 3 DtCOMPOfeBPi i NUMBEPi i STANDING- STAL-K.& 1L 0 £ Hi 0 £ £ IO II II 12 MONTHS Figure 8. Proposed scenario for the combined effect of decomposer and herbivore/ xylophile life histories on stalk quality plotted with the attrition pattern of the 1984 Agave palmeri age class. 26 LAP^&U CELL TUNNEL ambient TBMPePiArruP»e ^MAOBD- 40 Z£ IO:0O TIMS Figure 9. Temperature profile on 6 June, 1985 of Xylocopa californica arizonensis nest in a larval cell and in the nest tunnel. 27 plants with flowering stalks which are typically shorter and more slender than those of A. palmeri. Their life histories are also quite different from A. palmeri in that they have the capability to flower repeatedly. D. wheeleri and Y. elata wood is more dense than the inner wood of A. palmeri (Table 4). Also, these two woods are of a single or structurally homogeneous type throughout while A. palmeri flower scapes are composed of two distinct layers in cross section. The outer layer, regardless of stalk size, is approximately 6 mm. wide and consists of many fibers densely arranged so that they are difficult to penetrate, whereas the entire inner area is a very low density pith with many air lacunae similar to balsa wood. Table 4 includes values for only the inner layer of A. palmeri. Comparison of these three woods shows that as density increases the conductivity(k) increases (eq. 3). The insulation effect (inverse of k) decreases as density increases thus the very light core of the Agave plant has a comparatively high insulatory value as compared with the other woods listed. On the other hand, density may be a good predictor of life expectancy for standing flower scapes. D. wheeleri and Y. elata both tend to remain intact and connected to the plant for 2 or 3 years (unpubl. data). A complete analysis of the thermal environment of an exposed stalk would include morphological features of the stalk (e.g. surface texture, color, diameter, shape) and how these influence the thermal conduction, convection and reflectance of radiant and ambient heat, but this was not attempted here. Although the thermal conductivity values (Table 4) do clearly show that there is resistance to heat flow it appears that the importance of insulation is insignifigant due to the 28 Table 3. Breakdown of stalk conditions (explained in text) and carpenter bee usage in April, 1986 Stalk characteristics: N % holes Total stalks 78 100.0 15 1983 age class 3 3.8 1984 age class 15 19.2 1 total no. of stalks > 1 year old 18 23.0 1985 age class 61 78.2 condition 5 3 3.8 condition 6 58 74.3 south exposure 4 5. 1 1 west exposure 3 3.8 2 south and west exposure 12 15.4 2 unknown 3 3.8 stalks > 1 yr. old or exposed 57 73. 1 6 1 yr. old and shaded stalks 21 26.9 9 Carpenter bee usage stalks % total no. of with stalks nests nests total no. of stalks with nests 15 19.2 1984 age class stalks with active nests 1 0.01 2 1985 age class stalks with active nests 14 17.9 15 stalks shaded 10 12.8 exposed 5 5.1 5 south 1 1 west 2 2 both 2 2 29 small size of stalks. Rapid heat fluctuations of stalks exposed to sunlight (Fig. 9) indicated that stalks are too small in diameter to slow down heat transfer appreciably. Therefore in terms of the bee nest, maximum temperature that an exposed nest will reach is directly related to the air temperature. Although this factor may not prevent unshaded carpenter bee nests from being successful in climates with lower ambient high temperatures, it was a major factor at this study site. The rapid temperature response of exposed stalks suggested two predictions about carpenter bee nesting ecology. One, that stalk diameter above that needed to house the nest tunnel (15-20 mm) is not an important component in nest site selection and second, that reduced stalk temperatures can only be attained by shading. Stalk selection based on size was not apparent in this study. The one nest I found built in an exposed stalk was abandoned in May before brood could have matured. Substrate Availability Table 3 summarizes nest substrate availability within the Rincon Creek Ranch study area in April, 1986. This is a compilation of nest site conditions based on stalk age and thermal environment of the microsite which the preceding data have indicated are important factors toward successful nest establishment. When considered together these can provide an estimate of stalk suitability for Xylocopa. At this time, sixty percent of the standing stalks chosen by nesting females were in the 1 year age class. Of these 60 stalks, 45% (or 27) were unfavorable because of southern or western exposure. Therefore, the number of 30 Table 4. Thermal conductivity and density measurements for three wood types TYPE DENSITY CONDUCTIVITY(k ) R-VALUE g/cm3 BTU(ft°C hr)'* 1/k Dasilyron wheeleri 0.36 1.05* 0.95 Yucca elata 0.24 0.86 1.17 Agave palmeri thick 0.05 0.56 1.79 thin 0.05 Hardwoods (maple, oak)* 0.72 1.09 0.91 Softwoods (fir, pine)* 0.51 0.8 1.25 * Values from ASHRAE Standards Handbook 31 potential nest sites was 35 or roughly 2.5 stalks per hectare at this site. Based on these criteria the 1986 stalk population habitable by X. c. arizonensis can be estimated in two ways. The first gives a minimum estimate and is obtained by adding the number of stalks shaded in the appropriate directions (21) and the number of exposed stalks that were used for nesting (6) which was 27. The second estimate of substrate availablity is a maximum limit and included are all 61 standing 1- year- old stalks. This value is justified if shading by the leaves is adequate and no preference is exercised by the bees. Since only the area behind the leaves is shaded less volume per stalk is available for use if all stalks are included. These figures are rough estimates of the total substrate available. By knowing the amount used and the percentage of suitable stalks available an estimate of percentage of habitat used can be calculated. For 1986, the number of nests was 15 and there were 61, 1 year old stalks of which 21 were shaded appropriately. Fifteen nests divided by 21 suitable stalks gives 53% substrate usage and 15 nests divided by 61 1 year old stalks gives 25% substrate usage (Table 3). Substrate Usage Patterns Table 2 summarizes the microsite conditions of active nests and inclusive time duration the nests were used for 1985 and 1986 at the field study area. Most nests were excavated in April and May with a small resurgence of nest construction in August and September. Thirty two nests were recorded in spring of 1985 and 25 in spring of 1986 while only 13 appeared active in late summer of 1986. Three cases were found 32 of two nests built in one stalk. Thirty of the 31 active nests in 1985 were abandoned in spring of 1986 and 16 of 25 were abandoned between spring and summer 1986 (Table 5), indicating that nest movement is common, especially between years. Since only 4 new nests were found in the late summer 1986 censuses, it is possible that some individuals died or joined other nests. Nest abandonment or failure was common and occurred regularly between years and often within the same year. Nest abandonment censuses based on external appearance of the nest in spring and summer, 1986, are listed in Table 4. Only 362 of the 22 nests established in spring of 1986 were used in the fall and 8 of 16 nests categorized as abandoned had no external structural problem evident. Thus, many nests were used for only one brood even when the stalk itself remained standing and was externally structurally intact. Although available data are scanty, there are indications that vertebrate predator damage did not always result in nest abandonment. Both Gila Woodpecker (Melanerpes uropygialis and Gilded Flicker (Colaptes chrysoides) attack nests. In 1986, 5 of the 8 nests used in both spring and late summer nesting periods were damaged by birds in the spring season, yet later appearance of sawdust indicated new nest excavation and cell provisioning. Thus, external damage to nests did not always result in nest abandonment. This suggests that internal deterioration of the wood surrounding the nest burrow may be a primary cause of abandonment. Further evidence that nest abandonment results from biotic use of materials surrounding carpenter bee nests comes from the nests established in the laboratory. Here, nest abandonment was rare (Table 33 6). Overall 10 of 11 nests remained constantly occuppied until the final cencus in May, 1987. Ther single case of abandonment was in an observation nest made of redwood with a 5 x 5 x 25 cm. outside measurement. During this short length permitted only 5 cells to be provisioned which filled the nest except for 12 to 15 mm. on both sides of the entrance. It appears that the nest wall scraping necessary for cell partition building may have enlarged the nest to a point where further nesting would have been difficult. These data indicate nests are preferentially re-used, if the wood remains suitable. 34 Table 5. Factors Correlated with Nest Abandonment in Spring, 1986 no. nests abandoned from previous season 16 woodpecker damage 3 stalk fallen 5 usurped by spider 1 abandoned (reason not evident) 8 total 17 Table 6. Length of nest occupation in stalks kept in the laboratory Stalk Nest # Nest begun Months active abandoned? A 1 4-86 13 No 2 4-86 13 No 3 6-85 23 No 5 5-85 12 No B 1 5-85 12 No 2 8-84 33 No I 1 4-85 25 No 2 5-85 24 No H 1 4-84 37 No Observation Nest 14 2-86 12 Yes Discussion The life histories of the wood nesting carpenter bee and its nesting substrate are interactive. The ephemeral nature of the stalks and the variety of factors which cause their deterioration make the nesting substrate an unpredictable resource to the bees. The probability that a stalk will fall greatly increases in the summer following reproduction (Fig. 8), however, internal deterioration often precedes stalk failure. Flower scapes of A. palmeri that are suitable for nesting are available for only a short period. On the other hand, female X. c. callfornica probably live up to two years and reproduction may be delayed for 9 months after maturation. This is known for X. virglnica, (Gerling, Hurd and Hefetz, 1978) a member of the same subgenus (Xylocopoides), as X. c. arizonensis. Preliminary observations indicate the 2 species share these life history characteristic. Thus, a four- month window occurs when stalks are suitable for carpenter bee use. However, since stalk attrition is high during this period all stalks have an unpredictable life expectancy. The most predictable factor is that A. palmeri flower scapes will probably not last through another generation of bees. Therefore nests are abandoned permanently between generations or used through one summer and abandoned the following spring. X. c. arizonensis exploits the flower scapes of A. palmeri after flowering is completed. At this time in the plant's life cycle, energy stored in the bole has been expended on reproduction and both vegetative 35 36 and reproductive parts have usually dried. Since this plant's reproduction is synchronous over a 2 month period, stalk senescence is also coincident. Despite this homogeneity, the carpenter bees failed to use many stalks, indicative either that some selectivity did occur or that this population is not limited by nesting substrate availability. When A. palmeri stalks are categorized according to age (or quality) and thermal characteristics of their microsite, the data show signifigant nonrandom use by nesting X. c. arizonensis. Females utilized only a small percentage (19%) of the total available stalks standing at the site. However, females preferentially selected nest sites in 43% of those stalks that were structurally sound and shaded to the south and/or west. As shown in Fig. 5 the chance that a stalk will fall increases rapidly in the summer that follows flowering. Downed stalks are unsuitable for X. c. arizonensis and were found containing bees only during the winter months when all residents were adults and in diapause. Risks to fallen nests containing young may lie in increased exposure to the sun or perhaps increased vulnerability to ant predation. Nest abandonment while the stalks were intact and standing suggested that some agent was responsible for degrading the substrate prior to its falling. The stalk attrition curve supports the hypothesis that decomposer and herbivore/xylophile life histories interact additively to weaken the stalk. Few stalks fall until 12 to 16 months after blooming. This period begins midway through the summer following flowering and continues until October. Therefore, even though the hypothesis remains experimentally untested, observations indicate that the rate of stalk 37 attrition is closely associated with life cycles of associated biota. Carpenter bee nests are themselves a cause of stalk weakening. Although this is not a consistent cause of stalk fall, stalks are observed to break at the nest. Fowler (1983) has shown that a statistically signifigant number of fallen Yucca elata stalks contain nests of X. c. arizonensis. Saponins and flowering phenology exert positive effects on carpenter bee nest site availability. Saponins are present throughout the Agavacea and are a known insect deterrent. Their presence may best explain why no insect taxa are known associates of living stalks (Waring and Smith, 1986). Clearly, once dead, the soft wood of the A. palmeri stalk is readily colonized by a host of arthropod taxa. The anti- herbivore effect of saponins delays invertebrate colonization until late summer and therefore have little impact on first season stalks. Growth and colonization of herbivore/xylophiles follow the same pattern. In this, way, saponins and timing of flowering contribute directly to the availability of X. c. arizonensis nesting substrate. The absence of nests built in green stalks is initially discordant with the hypothesis that nest sites are constructed in stalks that are structurally stable. Living stalks were found to be superbly adapted to remain standing throughout the flowering period. In 1986, stalks fallen before flowering were observed three times: 2 cases felled by overhead tree branches and one which had grown from a pack rat nest (Neotoma albigula) and had been chewed off at the base, probably by the rodent. Observations of nesting in green stalks during the late nesting period, 1986, affirm that stalks can be used after flowering but before 38 drying. Certainly solitary bees do have problems with nest moisture since it promotes harmful microbial growth on provisions. The dufour gland secretions can provide a hydrophobic cell lining (Cane, 1981; Hefetz et. al, 1978; May, 1974). That this ability arose early in the evolution of the Apoidea is indicated by the ocurrence in the primitive Colletidae (Albans et. al., 1979) and this is known in the xylocopine genus Proxylocopa (Kronenberg and Hefetz, 1984) and others. Moisture in living tissues may likely deter bees from nesting in green stalks. Avoidance of "green-colored" stalks could also be due to biophysical constraints of the plant. Green stalks are smooth and this could prevent bees from landing or remaining stationary while digging. The two nests in green stalks seem to support this assertion. Both nests were started at a point where prolonged contact with a tree had damaged and roughened the surface. Nest Site Selection and Thermal Characteristics Internal temperature of unshaded stalks in the field regularly exceeded 43°C for extended periods. When blocked from the direct sunlight, stalk temperatures remained close to ambient. X. c. arizonensis females seem to prefer to build nests in stalk material located above the protection of the leaves but chose stalks shaded in the hot directions (south and west in the northern hemisphere). This selection reflects avoidance of the greatest heat load of the day. Shade was most often provided by an adjacent perennial plant although nests were also located behind the dried basal leaves of the Agave plant itself. Nest usage patterns suggest that temperatures attained in 39 exposed stalks may exceed thermal maxima for this species and missplaced nests would destroy any progeny of females who choose exposed nest sites; therefore, strong selection must operate on female nest site choice. Carpenter bees are active during the heat of the day (Gerling, Hurd and Hefetz, 1983; Chappel, 1982; Louw and Nicholson, 1983; Nicholson and Louwe, 1982) which has stimulated study of adult thermal biology for the sister species X. v. virginica (Baird, 1986), another subspecies, X. c. callfornica, (Chappel, 1982), as well as a sympatric species, X. varipuncta, (Heinrich and Buchmann, 1986). The latter two species were found to tolerate temperatures over 45 °C for short periods. The mechanism for cooling in these two species appears to be mainly convective transfer enhanced by the large frontal area of the head and rapid flight. However adult activity and preadult cooling mechanisms are not directly comparable. As is true for all the bees and wasps the egg, larval and pupal stages are immobile in the natal cell making adult behavioral mechanisms such as relocation, increased flight speed or evaporative cooling unusable. Therefore, preadult stages are reliant on appropriate site choice by the parents for protection from lethally high temperatures. These bees appear to distinguish between and prefer stalks with partial cover from adjacent vegetation even though virtually all standing stalks are shaded at the base by dried leaves. The data strongly suggest that successful habitation of stalks in the Sonoran Desert is partially determined by amelioration of radiant energy. The exclusive use of those regions of the stalks shaded by adjacent vegetation or by the leaves of the plant itself indicates that 40 full exposure to the sun is detrimental to X. c. arizonensis immatures. This study also corroborates the observed trend that nest site location varies according to ambient temperatures. The mechanism that females use to discern the thermal characteristics of stalks is not known. The only two observations of females initiating nests have been made in the late afternoon. If these two observations are representative, that is if all nest building begins in the afternoon, this would provide a simple and effective behavioral mechanism for the bees to evaluate the microclimate of potential nesting sites. Also, digging through the outer layer of the stalk takes 2 to 3 days and until the nest is sufficiently deep for the female to completely enter she digs only during the day, returning at night to her natal nest (Minckley, unpubl.). During this time evaluation of the thermal microenvironment is also possible; however the decisions appear to be made quickly since shallow nest cups, indicating nest initiation and subsequent abandonment, were extremely rare at this site. The stalks themselves and the thermal environment appear to present a heterogeneous resource to the Xylocopa which the bees then must evaluate. A. palmeri are patchily distributed in the study area and appropriate nesting sites are a subpopulation of the total stalks present. Additionally, as the declination of the sun changes, the shadow cast by adjacent plants will change along with the normal seasonal temperature cycles. A newly nesting female is confronted with the task of locating stalks, determining the structural quality of the wood and assessing thermal regimes of the microenvironment. Variation in all these factors creates a complex heterogeneous environment. The interaction of the environment and the bee would appear to encourage 41 selection for some plasticity in the behaviors associated with evaluating nest site suitability. Active choice and not a passive mechanism is surely reponsible for location of those specific sites that are suitable for rearing young. The value of selecting an appropriate nest site is great. Compared to most insects carpenter bees have a low fecundity. They also exhibit parental care (Gerling, Hurd and Hefetz, 1979; Michener, 1972). Gerling, Hurd and Hefetz (1978) estimate the lifetime reproductive output of X. v. virginica to be from 6 to 14 eggs, and 14 cells were the maximum found in one provisioned X. c. arizonensis nest (Buchmann and Minckley, unpublished). Females remain in their nests to guard the young during the 45-60 day development period. Immatures develop in a natal cell and at maturation remain unable to fly for another 2 weeks (Gerling, Hurd and Hefetz, 1983; pers. obs.; Skaife, 1952). The females' selection of a nest site must therefore remain suitable throughout this period. The basis of nest site selection in X. c. arizonensis are factors which directly affect individual fitness. Incorrect assessment of structural quality of the stalk or of the thermal microenvironment can result in complete nest failure and death to all or most of a females' progeny. Thus, natural selection acts quickly through potentially severe differential reproductive success. It is clear that strong selection pressure exists for correct assessment and choice of a nest site. 42 Limiting Factors on Nesting Success From this study, nesting success appears to be derived from the appropriate evaluation of substrate age and thermal environment by nest- initiating females. Effects of parasites also negatively affect population size and may provide an alternative explanation to nest usage patterns observed in this study. The only known parasite of this carpenter bee is the beefly Anthrax slmson habrosus, a bombyliid fly (Hurd, 1959; Marston, 1970). Parasitism of X. varipuncta nests used for several years do appear to be more heavily parasitized than first year nests (Minckley, In prep.); however, this species differs importantly from X. c. arlzonensis by continuously inhabiting nests for years. Nest re-utilization should facilitate subsequent infestation by the emerging beeflies. Field collected X. c. arizonensis nests rarely contain parasitoids and are never found to be parasitized at high levels. The high rate of nest abandonment and short nest occupancy found in X. c. arlzonensis probably functionally decreases parasite infestations and, although patterns of nest abandonment observed in this study do correlate with a parasite escape hypothesis, I believe that the lower parasitism rate is a secondary benefit and that the ephemeral nature of the nesting substrate is the primary influence. In this study it is not conclusive whether the availability of nesting substrate limits population size since only 25% to 53% of the stalks which fit the substrate quality and thermal criteria of sites were used for nesting. Parasites, which are known to inflict considerable damage on some solitary bee nesting aggregations (Linsley and MacSwain, 1952), appear to have little effect on X. c. arizonensis. 43 A problem with this type of study 1s defining suitability accurately. For example, thermal criteria of stalks in the field are only rough estimates of the actual conditions at the site although I have tried to make my suitability criteria falsifiable. Another problem is that short 9 term studies of populations easily underestimate long term effects. In this study, few A. palmeri may reproduce in some years and cause nest site availability to become limiting. If so, the low fecundity of carpenter bees might prevent a rapid response to the population crash. Smith and Whitford (1978) have proposed that pollen crop at some sites is limiting for this species from evidence which compared pollen availability vs. percent habitation of trap nests at various sites. I feel that this may be a tenuous comparison, however, since trap nesting appears to be a method which does not accurately estimate carpenter bee population size (Krombein, 1967). Speciation in the Xylocopini The selection of an appropriate nesting substrate and site is highly developed in the Xylocopini as pointed out by Hurd and Moure (1963): "Superficially it would appear that the selection of nesting wood by bees of the genera Lestis and Xylocopa is largely indiscriminate. However, if the character of the nesting woods is considered, rather than its botanical affinity, it is apparent that a relatively high degree of preferential selection is exercised by many of the species" My study demonstrates that female carpenter bees do evaluate nesting substrata, even though at the Rincon Creek study site only one 44 species of plant Is available for use. Choices made by this population were based on features of the environment which can adversely affect reproductive success. These are substrate quality and the thermal microenvironment of the nest. Nest site selection has been discussed by Hurd and Moure as a mechanism for isolation of subpopulations and eventual speciation in Xylocopa: "It is evident that predilection for a particular nesting substrate, though restrictive, does not preclude pioneering individuals-and subsequently local populations-from selecting nesting substrates new in character to the species as a whole. Our studies suggest that this process is essential to the evolution of new forms." (Hurd and Moure, 1963) It is apparent that in this entire subfamily of wood nesting bees, discrimination of wood as a nesting substrate is a primitive characteristic that was instrumental in the early differentiation from other anthophorids (Hurd and Moure, 1963). Species have become more fine tuned and usually discriminate substrate based on some quality (e.g. density) or in a few instances plant types. For example, X. hottentota utilizes only flower scapes of Aloe and the entire subgenus Biluna specializes mainly in bamboo (Hurd and Moure, 1963). However, this is a sympatric speciation argument analogous to that proposed for tephritid flies by Bush (1974) and therefore controversial despite a substantial amount of theoretical work which supports existence of this speciation mechanism. Theoretically, it has been shown that populations which inhabit heterogeneous environments can maintain stable genetic polymorphisms and eventually speciate (Seger, 1985; Maynard Smith, 1966). Isolation by divergent habitat preference 45 can lead to genetic differentiation if mating is assortative. Maynard Smith (1966) points out that this speciation scenario is possible even if two habitats differ in quality. Isolation in this case is perpetuated by the selective advantage accrued by experience in the natal environment which outweighs the chance taken in recolonizing another habitat. Since Xylocopa appear to tend toward the use of nesting substrates similar to those they were born in, experience may be the precursor that eventually allows isolation. My research demonstrates that nest site selection in X. c. arizonensis is based on environmental factors and it is variable among individuals within a population. Although confined by fitness constraints this behavioral trait is labile and not rigidly controlled genetically. For example, in the field one nest, although unsuccessful, was constructed in full sun. Female selection of nesting substrates was also not based on wood density. Density of the three nest woods tested in this study varied widely (Table 2) and nests are commonly found in all three. Novel nest substrata are also used in urban Tucson and nests are found both in redwood fencing and the hollow culms of Phragmites communis among others. Plasticity in occupation times also occurred. In the field, one nest was occuppied for two years and in the laboratory one was occuppied continuously for three years. On the other hand some nest substrate preference is apparent. I have not found nests in large tree branches, although they are reported to occur in some tree species (see Table 1) and are frequently used by another species in this area, X. varipuncta. Also, this species is difficult to induce into artificial trap nests (Krombein, 1967). Whitford and Smith's (1979) success rate was extremely 46 low using this technique for X. c. arlzonensis, indicating some avoidance even though a hole is pre-drilled and thus should take little effort for a bee to extend and make habitable. Plasticity of nest site selection behavior is obviously adaptive for this species. Nest substrate quality and thermal microenvironments are highly variable within this one study site and without doubt are more variable over the entire range of this species. An ability to analyze each site and decide whether it will remain intact and within the adequate temperature range directly affects fitness. Natal experience may be highly influential in this system. Genes which enhance the ability to remember the natal nest site would confer an advantage to appropriate nest site choices made in the next generation by increasing the probability the nest will also be in a suitable site and substrate. Behavioral plasticity seen in X. c. arizonensis may in fact be maintained as an evolutionary response to fluctuating environments (Levin, 1968; Maynard Smith, 1966). This alone is not sufficient to allow speciation to proceed since assortative mating is also necessary. Some evidence suggests that assortative mating does occur in this species (Fowler, 1983; pers. obs.). Male territories are at nests as well as at food resources and males will establish territories close to their natal nest. Fowler (1983) marked 11 males in the winter and located 6 of these with territories in the same area the subsequent spring. Therefore, the requirements for a sympatric-type of speciation mechanism to occur are present. Although some selectivity in nest substrate choice was 47 demonstrated in the present study it is difficult to apply these data to a definitive categorizaton of nesting preference for this species, as discussed by Hurd (1956). Many types of woods are used which represent a wide range of densities, shapes and sizes. There is high variance in the nesting substrata used by this species and nest site selection may thus be understood in terms of what is availabile in the area, its structural integrity and thermal microenvironment. Because of this, I feel that the isolation which led to differentiation most probably occurred through geographic isolation from either X. c. californica or X. virglnica. I should point out that Hurd (1978) or Hurd and Moure (1963) did not postulate that X. c. arlzonensis diverged by a sympatric speciation process, but only asserted that this may have occurred in the Xylocopa. My data may thus not apply to this particular problem. However, using the data I have collected it is difficult to envisage genetic differentiation occurring by isolation due to differences in nest substrate selectivity. Other Xylocopa, however, may express more specificity in their nest substrate selection. Since the genus Xylocopa is most speciose in the tropics perhaps those species should be studied more intensively for evidence of this mechanism. Summary The data presented strongly suggest that female X. c. arizonensis excercise active choice in their selection of a nesting site. Two factors appear to be primary determinants of site selection: structural condition of the substrate and the thermal microenvironment. Females respond to the ephemerality of the nesting substrate by occupying nests. CHAPTER 2 LIFE HISTORY, NESTING BIOLOGY AND MALE MATING BEHAVIOR OF THE LARGE CARPENTER BEE, XYLOCOPA (NEOXYLOCOPA) VARIPUNCTA (HYMENOPTERA: ANTHOPHORIDAE) Introduction Xylocopa varipuncta is a large carpenter bee (Anthophoridae: Xylocopini) and the northernmost member of the subgenus Neoxylocopa. This subgenus has the greatest number of species (47) and widest distribution of any of the New World subgenera (Hurd, 1973; Hurd and Moure, 1963). They are, however, largely understudied. All Neoxylocopa species are sexually dimorphic with black, females and yellow or tawny males. Recently an exocrine gland has been described and chemically evaluated that is also a sexual dimorphism from males' of two Neoxylocopa, X. gualensls (Frankie, Vinson, and Williams, 1986) and X. varipuncta (Andersen et.al., 1988). The function of the three secretory products has not been determined by bioassay but field studies suggest it is likely a female sex attractant (Andersen et. al., 1988). Neoxylocopa species have recieved few biological studies due to taxonomic confusion from the dissimilar sexes (Hurd and Moure, 1963; Hurd, 1978), minimal occurrence within the United States and difficulty in finding nests. From the biological data available it appears that at least some of their life history characteristics are common to other Xylocopa. These include their foraging habits, adult longevity and social organization. They have highly generalistic 48 49 (polylectic) foraging utilizing both exotic and indigenous plants (Barrows, 1980; Gerling, 1984; Hurd and Linsley, 1975). Females live up to 2 years and once mature appear to have few natural enemies although there is mortality in the immature stadia. The primary natural enemy of X. varipuncta appears to be the bee fly Anthrax simson habrosus (Marston) in the family Bombyliidae which attacks mature larvae (Minckley, in prep.). Female lifespans overlap those of their progeny, trophollactic feeding occurs between all adults in the nest, and division of labor occurs. These are the conditions that define sociality in bees (Michener, 1974). A solitary existence is also possible as these same bees can develop and survive without the assistance of other adults. This level of sociality is thus weak and appears to represent a transitional state between solitary and true eusociality that has been labeled "metasociality" by Velthius and Gerling (1983). Males, however, live ca. 10 months and are expelled from their natal nests during or prior to the mating season. Male mating behavior for all known Xylocopa spp. involves use and aggressive defense of territories. Two general tactics are employed which differ primarily in the males' role in mate choice. In one type referred to as resource defense territoriality (Emlen and Oring, 1977; Bradbury and Vehrencamp, 1977), males hover or patrol for mates at flowering plants and/or at nests. This type of behavior is described in detail for two members of the subgenus, Xylocopoides, X. californica californica (Cruden, 1966) and X. virginica virginica (Barrows, 1983) and is not known from any Neoxylocopa. Territoriality of this type results in male mate choice since females appear unable to fend off male copulatory attempts and are 50 harassed, whether mated or not, any time they enter a males' territory (Barrows, 1983, pers. obs.). The second type of male mating tactic used by Xylocopa has been referred to as male dominance polygyny (Emlen and Oring, 1977), non- resource defense (Alcock, et.al., 1978) or lek territoriality. X. varipuncta, X. (Neoxylocopa) hirutissima and X.(Koportosoma) pubescens are three Xylocopa spp. whose males use this reproductive tactic. Male territories are located in plants or under building overhangs (Gerling, Hurd and Hefetz, 1983; Linsley, 1965; Marshall and Alcock, 1981; Velthius and deCamargo, 1977). Most territories occur at sites without nectar and pollen food resources to attract females, that is, they are not associated flowers plants and contain no nests. Observations of females within male territories are rare but resident males make no attempt to chase or grasp females (Alcock, 1987; Velthius and deCamargo, 1977) unless copulation ensues. These observations indicate that mate choice is solely made by females. Male mating behavior in X. varipuncta is among the most completely studied for any Xylocopa species (Alcock and Smith, 1986; Marshall and Alcock, 1983; Minckley and Buchmann, in prep; this study). The mating tactic used has been called "lek territoriality" (Marshall and Alcock, 1981) because males do not contribute to parental care of their progeny and their territories occur in non-flowering trees and bushes located away from nests. Since female-attracting resources are generally not within territories it is not immediately evident what mechanisms are used by males to attract females. Each study on this species has addressed this topic and has proposed either a pheromonal 51 cue (Andersen et.al., 1988; Alcock and Smith, 1986; Marshall and Alcock, 1981; Minckley, and Buchmann, in prep.), visual cues (Marshall and Alcock, 1981) or strategic location of the territory site (Alcock and Smith, 1986). This paper presents data on the nesting biology, life history and male mating behaviors of X. varipuncta. In addition to data on male mating tactics with emphasis on the mechanisms used to attract females, information on nesting and life history are important and are included because of the paucity of data available on these aspects of X. varipuncta biology, and recognition that many aspects of a species' ecology can affect their social and reproductive behaviors (see Wrangham and Rubenstein, 1986). Description of the study site The study site, called herein the Ninth St. site, was 4000 meters square area with a a group of 5 chinaberry or Texas umbrella trees (Melia azedarach L.) which contained between 25 and 30 nests of X. varipuncta on its' south side (Fig. 10). These trees were in the frontyard of a residence at 729 E. Ninth St, Tucson, Az., 1 km. southwest of the University of Arizona. Perennial vegetation was both exotic and indigenous to the Sonoran desert. Topograhically the area was flat with the most noticeable relief associated with a small wash located ca. 0.5 km. south of the site. Materials and Methods General Activity Patterns and Life History Female activity patterns were determined by direct observation of bees exiting and entering nests. To calculate the number of foraging trips (In minutes) entrance and exit times were recorded and duration of the observation period noted. Females also were categorized depending on presence or absence of pollen in their scopae. Events which occurred Inside nests were determined indirectly by positioning styrofoam drinking cups below 5 nest entrances. In this position, the cups collected discarded nest refuse; feces, exuviae, wooden partition material, sawdust and nest parasites. This method allowed the timing of otherwise unobservable life history events within nests. Temperature data were collected in 1985 with an Omnidata Easy Logger Recorder and in 1984 and 1986 from the U.S. Weather Service station at the University of Arizona. Male Mating Behavior Male behaviors at territories were recorded on a Realistic Micro* 20 mlcrocassette recorder or noted on paper when observed incidental to other types of observations. Time budget data was calculated by focal method observation of a single individual for 15 min. or until they left (Altmann, 1973). The four behaviors which occurred during collection of these data were, (1) hovering, (2) male-male interactions, (3) marking or grooming and (4) male-female interactions. Further details were 52 53 determined from video-tape records of several displaying males at another location in 1987. Data on display site usage was recorded by walking a 200m transect every 0.25 hours. When a territorial male was located, the plant species that contained the focal point of the territory (referred to hereafter as the focal plant), location, flowering status, height above ground the male hovered, and compass orientation of the territory (measured with respect to the central axis of the tree) were recorded. I found that different positions were often used on one plant so I distinguished between a "site" and a "station". Site was defined as the individual plant used and station as the position (height, orientation) in the plant. All plants along the transect were mapped and their heights determined. Results Nest Substrate Usage and Nest Architecture Table 7 is a synopsis of my observations and published reports of the nesting substrates used by X. varipuncta. Life History Life history data is summarized in Figure 11, and for discussion I begin with onset of activity in the spring. Overwintering ceases in late February or early March after 4 consequetive days of ambient temperatures in excess of 20°C. This was determined by correlating first activity for three years with temperature data for the previous month (Fig. 12). These dates were March 1, 1984, February 16, 1985 and February 19, 1986. Bee flight/ foraging did not occur daily but was observed on any day that temperatures exceeeded 20-21 °C. Mating activity, discussed in depth subsequently, began 2-3 weeks after termination of winter quiescence, and lasted 5-8 weeks. Males remain in nests along with females until mid-April. Thereafter, females bodily blocked nest entrances or physically pushed out all males attempting to enter nests. Nearly coincident with male expulsion, the appearance of fresh sawdust in nest cups indicated that tunnel construction had begun. Female activity outside nests was higher than any other time of the year (Fig. 13) based on return trips per unit time, and the high percentage of return trips with pollen indicated that larval cells were actively being provisioned (Fig. 13). Sawdust often appeared irregularly from single nest entrances, punctuated by intervals of several days, suggesting that all females may not be synchronous in their nesting activity. The appearance in June of feces, partition material, dead or parasitized larvae, or pupae and exuviae of the bee fly A. s. habrosus in nest cups signaled that partitions had been destroyed and the larvae had developed. Egg-to-adult metamorphosis thus took 45-55 days. After this period, large quantities of adult feces were collected in nest cups or collected on wood below entrances apparently from teneral adults which had not yet flown. Active adults defecate while in flight (pers. obs.). Other than feces, nest cups collected little material during the remainder of the summer. The rate of daily female foraging trips also declined dramatically as did the number of females observed carrying pollen (Fig. 13). This suggests that no other cells were provisioned and a second generation was not reared. Table 7. Nesting substrata used by X. varlpuncta Plant species Citation 1. Mesquite this study Prosopis .lullflora 2. Chinaberry or Texas umbrella this study Melia azedarach 3. Elderberry this study 4. Ponderosa Pine this study Pinus ponderosa 5. Cottonwood Hurd 1978; this study Populus fremontli 6. Palmers agave this study Agave palmerl 7. Pepper tree Hurd 1978; this study 8. Alder Hurd 1978 Alnus oblongifolia 9. Apple Hurd 1978 Malus 10. Apricot Hurd 1978 Prunus armeniaca 11. Balsa Hurd 1978 Ochroma lagopus 12. Chinese paper plant, oak Hurd 1978 Quercus agrlfolla 13. Organ pipe cactus Hurd 1959, 1978 Cereus thurberi 14. Eucalyptus Hurd 1978 15. Walnut Hurd 1978 Juglans major 16. Oleander Hurd 1978; this study Nerium oleander 17. Yucca Hurd 1978 18. Structural wood Hurd 1978; this study 19. Driftwood Janzen 1964 56 NWWWWM !-EX.I L&Sr&NP o UNU&EP PL-fikNT& o » NEST PLANTS Figure 10 Map of study area showing lek and nest sites. Refer to figure 5 for letter designations. 57 As stated previously, male mating behavior occurred 2-3 weeks after quiescent overwintering individuals became active. Male activity commenced daily between 1530 and 1630 MST and usually lasted until one- half hour before sunset. Thus early in the season males quit displaying between 1730 and 1800 h. while later in the season they often displayed until at least 1900 h. Behavior at Territories Male display behaviors. Time budget data for territorial males predominated with behaviors I associate with male "display". Three and one half hours of focal individual observations on 14 males at territories between March 21 and April 5, 1985 and March 10 and April 4, 1986 at the Ninth Street site. Males spent 82% of the time in display behaviors, 16% in male-male interactions and 2% or less engaged in grooming or scent-marking behavior or female-male interactions. No matings were observed during this time period. Display behaviors consisted mostly of 1-12 second flights in loops or figure eights 1 to 2 m in diameter interrupted by 0.5 to 3 sec stops near the focal point. I call the area where this behavior occurs the "focal area". A focal point is usually a leaf or twig at the center of greatest male activity and serves as a platform for matings to occur on (Marshall and Alcock, 1981). Data from video tape revealed that males averaged 14 looping flights per minute each lasting an average of 2.7 s. (n=95; 0.6 to 11.5 seconds). Flights all stopped at the focal point and these "hovers" averaged 1.37 seconds. (n=99; 0.46 - 3.08 seconds). Thus while undisturbed, one male spends 32% of his time 58 hovering motionless next to his focal point and 68% of his time in these quick, darting flights. Hale-male interactions. The presence of another male at the territory initiated a range of defensive behaviors which differed in their levels of aggression. I've grouped these into two categories for descriptive purposes, but, oftentimes both occurred during a single confrontation. Vhich behavior the resident males used appeared to be determined by how the intruder male approached the residents' territory. The first behavioral type was typified by moderate to low level aggresslvity. Bouts usually began and were first noticed as an aggressor flew 1-2 m over the focal area in large (5 to 15 meters in .diameter) loops. Next, resident males flew up to the level of the aggressor where he also proceeded to fly in loops. Often another male(s) would also appear in the territory. All males would usually fly in the same general direction with similar flight paths. These flights were centered over the attacked males' territory and often lasted from 3-5 minutes. Beside their long duration they were curious because males' did not dart towards one another and physical contact was never observed. The second type of territorial dispute was clearly more aggressive in nature, generally lasted less time, and when resolved could end with two different possible outcomes. Fights began when the aggressor male flew directly into the focal area of the territorial male. The territory owner then responded by flying within 15 cm. of the intruder as if undergoing a quick investigation and quickly thereafter an attack. Males flew directly at one another and it was in these bouts that infrequent contact sounds (such as wings clashing) were heard. 59 Loops away from the focal area were short and usually remained within a few meters of the focal area. Encounters of this intensity rarely lasted more than 30 seconds to 1 minute with three possible outcomes. First, a winner could emerge who again began to display while the other male left, secondly, the interaction could continue but with less aggressiveness resulting in behavior much like the first interaction type described above. Thirdly, both males might display in the same plant and tolerate each others' presence. Although the first two outcomes were most commonly observed, if both males displayed the resident remained much closer to his focal point than usual and always faced the newcomer. Table 8 compiles data on 15 male-male encounters, whether they are type 1 or type 2 aggression types and the outcomes of these battles. There are 8 type 1 battles in which 7 ended with the aggressor leaving and 1 instance of both males displaying together. The 6 recorded type 2 aggresive encounters resulted in two cases of the aggressor leaving, two of 2 males displaying and once of territory owner displacement. Which male won, aggressor or resident, was decided by whether the focal point the winning male used was the same as prior to the interaction. If it was, the male was considered a the returning resident and if not the owner was considered to be displaced. Female-male interactions. When females are present at territories, male behavior was not as variable. I observed two types of male-female interactions, female investigation and copulation. Female investigation of males in focal point areas is by far the most common and is so-named because females enter and hover in male focal areas, i i 1 1 1 1 1 1 1 1 SEASON ACTIVITY MALE TEPifMTOP.IAL.rTY ' NEW WEST CONSTRUCTION WOOD ^HAVINGrS . PPOS/fe|ONINGr/ EGGLATINSr FEcee/ pt»iA-£>£» &EEFUY PAPA'S 1 TBS I 1 1 I 1 1 ' I ' -J 1 2. =» 4. ^ 7 Q 9 IO 11 12- MONTHS Figure 11 Life history of Xylocopa varipuncta. 61 ! IZ 14 16. 2. 3 4- 5 <2» re&PiuA^-v* MAPiCH Figure 12 Temperature data for month preceding start of seasonal activity for 1984, 1985 and 1986. 62 MINUTB& MONTH Figure 13 Rate of female return flights and percent of trips with pollen. 63 within 20 cm of the resident male, and then depart. Each time one of these interactions vas observed a previously unobserved female flew into the territory and hovered 10-20 cm from the male. While there she constantly faced him until departing which invariably was less than 30 seconds each. Upon her arrival the male would do one of two behavioral sequences, 1) move quickly to alongside his focal point, facing her at all times while hovering motionlessly or 2) perform several quick "land and walk" sequences on the focal point and then hover alongside the focal point. "Land and walk" is a descriptive term first coined by Marshall and Alcock (1978). In this behavior the male lands at the bottom of the focal point, then rapidly walks to the top and flies off. Normally the male will repeat this action 3-5 times. After landing and walking, males always resume hovering alongside their focal point. Alcock (1986) observed, as did I, that observed matings are extremely rare. In his studies, only 5 matings were observed over 125 hours. In three years (1984-1986) only a single copulation was observed on 14 April, 1984. Copulation was in progress when the pair was first observed on a leaf of the ornamental pistachio tree (site B, station Bl) at the Ninth Street study site. From the time of first observation to their uncoupling was only 20 seconds. Display site selection In 1984 and 1985, 14 and 7 individual plants or sites, respectively, were used by males. Many sites were used in both years so only 9 unique sites were used in the study area. At these 64 Table 8. Types of male-male aggressive encounters and outcomes Date Aggression type Outcome 20-III-1984 1 aggressor leaves 2 2 males display 1 aggressor leaves 2 2 males display 29-III-1984 1 aggressor leaves 1 aggressor leaves ? 2 males display 1 2 males display 2 displacement 2 aggressor leaves 3-IV-1984 1 aggressor leaves 5-IV-1984 1 aggressor leaves 12-IV-1984 1 aggressor leaves 2 aggressor leaves 13-IV-1984 2 aggressor leaves 65 sites I observed 50 displaying males which occuppied 17 stations. No pattern was evident from the plant species used, their relative abundance or general appearance. The mesquite was the only native Sonoran desert species. No species was clearly preferred and plants used by males were a wide range of heights, widths and canopy densities. In the first two years of this study (1984 and 1985), it did appear that male X. varipuncta avoided resource bearing sites and thereby were in accordance with Bradburys' (1975) criteria that no resources are present and defended in lek territorial species. Twentyfour sites and ten sites were occuppied in 1984 (Fig. 14) and 1985, repectively, which did not contain nests or flowering plants. During the 1984 season, males appeared to discontinue use of plants that came into flower during male mating season. During this year, however, there were 7 instances when males did hover at resource-containing sites and 6 occurred in one chinaberry tree which was both in flower and contained three active nests in its' trunk. Twice at this site I observed females foraging within 1 meter of a displaying male. Yet, other than a quick investigation no apparent response from the male was seen This may implicate that a female chemical signal communicates her sexual status. The pattern of resource avoidance changed drastically in 1986. Although other experiments precluded gathering of enough data, it was clear that the preferred site was a pyracantha tree which flowered 4 out of the 6 week mating season. Spatial Patterns As suggested by Marshall and Alcock (1983) and Minckley and 66 Buchmann (in prep.) male orientation at display sites exhibits a strong preference towards the east-facing or central area of the focal plant. Twenty-five display sites were oriented eastward and 13 centrally compared with only 2 to the west and 3 to the southeast. One site was used most frequently in both 1985 and 1986, station B in 1984 and station E in 1985. Although the sites were within different plants, and disjunct by 18 meters both were along the property line that bisected the two houses nearest the X. varipuncta nest (Fig. 10). Males were conspicuously absent along other property lines, which tended to be as heavily vegetated, yet which had plants otherwise suitable for male mating activity. Figure 15 shows the frequency relationship between the number of displaying males and display duration time estimated from the presence/ absence data gathered during transects. Twenty-nine of 54 territories monitored were occuppied for less than 30 minutes. Sites occuppied for greater than 60 minutes represent only 20% (n=ll) of the total number. Short-term use sites were often the least-used sites throughout a season and those sites used for the longest periods were the most frequently used sites. Numbers of sites used per day combined for both years show that on 11 days 1 site was used, on 6 days 2 sites were used on 3 days 4 sites were used and 1 day 4 sites were used (Fig. 16). Single site usage was therefore most common and compatable with the dispersed lek territorial designation given X. varipuncta by Marshall and Alcock. (1978). Contained within the study area was a single site used more often 67 than any other during all years. Individual site and station usage data is shown for 1984 (Fig. 17). In 1984, the primary site was an ornamental pistachio tree which was used 2.5 times more often than the second most used site. In 1985 a pyracantha tree was used 3 times more often than the second most used site. Male preference for these sites was therefore strong and is further supported by the station data which shows that in 1984 station B, located in the ornamental pistachio tree, was used more than 4 times more than any other (Fig. 17). Figure 18 is a frequency diagram of the most used sites plotted against site usage for all other sites combined and is strong evidence for site dominance. I have not included the chinaberry tree station (E) because of physical separation by houses between it and the other sites. Overall, use of preferred sites exceeds all others in total numbers of days used and were used most often when only a single site was used. The numbers are the two sites were used 16 of 25 days that any male mating behavior was observed and on 7 of the 10 days total that I saw only a single male displaying. Third and perhaps most persuasive for the dominance hypothesis, these sites were used on 9 of 10 days I observed more than one male displayed. Besides indicating that males had a strong site preference this suggests that other males are attracted to the primary site either by the presence of the owner, some attribute of resident male or by the site location. Discussion Life History, Nesting Biology and Nest Architecture X. varlpunctas* preference for large nesting substrates results in extra material available for use by subsequent generations. This permits nesting to continue for multiple generations at the same location and facilitates the cohabitation of many individuals. Females thus have an option to remain in their natal nest dig new branch tunnels or disperse and search for another suitable site. Because of this, genetic relatedness within one nest is probably high since it is the daughters vho are most likely to inherit the nest. Guards limit access to nests (pers. obs.) and further increases the liklihood that close relatives occupy a single nest. This contrasts sharply with X. c. arizonensis which tends to use smaller, more short-lived woods that last only 1-2 years. Their nests are simpler, having only 1 to 3 branches and usually contain no more than a single female and her progeny. In comparison, the preference X. varlpuncta has for larger diameter nesting materials results in more complex nest architecture, increased nest longevity, and a higher level of social organization. The life history of X. varlpuncta is unusual because it is univoltine. X. pubescens (studied extensively in Israel) is similar in many respects to X. varlpuncta. Sexes are color dimorphic, males possess 68 69 a dorsal mesosomal gland (A. Hefetz, pers. comm.), they Inhabit large substrates for nesting, and nests contain many females. Both species live in desert climates, however in Israel, X. pubescens has up to 4 generations per year and male mating activity is nearly continuous (Gerling, Hurd and Hefetz, 1983). This difference is difficult to explain, but, perhaps is related to the nectar and pollen dearth which occurs in the Sonoran desert during June and July (O'Neal and Waller, 1984). X. varipuncta is a highly k-selected species as are all Xylocopa. The life cycle minimizes time in preadult stadia and maximizes time of adulthood. Development is continuous once the egg is laid and no preadult diapause occurs (pers. obs.) therefore prereproductive adults are present in nests for 9 months (June to March). Since females are long-lived, generations overlap and they are able to contribute parental care by trophyllactic feeding and nest guarding. Nests opened in winter have rarely housed females with extremely worn wings, suggestive that females die the same summer in which their progeny mature. Male Mating Behavior Seasonal activity began 2-4 weeks after the first individuals broke winter quiescence and varied due to persistence of inclement weather. Male hovering began daily between 1530 and 1630 h but ceased from 1/2 h before to just at sunset. Since time of sunset moved back as the season progressed, hovering males remained at their territories from between 1800 to 1830 h. in the early season and to 1900 h. and past as the season ended. Weather did affect daily mating activity, however, 70 ^ <£<* (S) Palo verde ~C F _Q) O Chinaberry-C E e .Q »<£ •o Figure 14 1984 site and station use showing flowering phenology of focal plants. 71 once the mating season was underway males were less reluctant to remain in their nests since I did see them hovering in relatively strong winds and at temperatures at 18° C, conditions which were unacceptable for activity several weeks earlier as the winter quiescence period neared an end. Male-female interactions at territories were, except for a single copulation, entirely of females investigating territorial males. This skew towards female investigative behaviors is consistent with reports by both Marshall and Alcock (1981) and Alcock and Smith (1986) and supports their belief that it is the females who control mate choice. In my observations of males confronted by females, the male consistently remained passive and close to his focal point, awaiting the female to land. Females always left unmolested and males never made any attempts to physically harass females. My observations are consistent with the idea of female "choosiness" and that the females are somehow evaluating and comparing males. This latter statement is inferential yet my observations are consistent because females at leks investigated and left more often than they were seen to mate. Alcock and Smith (1986) presented additional evidence. They observed females hovering by several males and copulating with one at plants where 2-8 males simultaneously displayed! From my data and those of others it is clear that male X. varipuncta should compete strongly for matings because there are so few opportunities to do so. Since females also probably mate just once (Marshall and Alcock, 1981) which my observations support, this further intensifies selection. Strong intrasexual competition of this sort would 72 MINUTES? Figure 15 Frequency of male display times for 53 males from 1984 and 1985 mating seasons. 73 be expected to enhance any characteristics which could positively effect mating success. In X. varipuncta these characteristics are expressed both in male behavior and morphology. Behaviors which advertise male presence, and quality, I have termed display behaviors. These consist of quick, elliptically-shaped flights punctuated by brief stops at the focal point. The purpose of these behaviors is conveyance of signals allowing females to locate males. It also appears there are two signal types, one visual and the other chemical. Visual signalling is accomplished by the conspicuousness of the males' golden-yellowish integument which contrasts markedly with female color. Although sex discrimination by color alone has not been demonstrated for female X. varipuncta, it has been shown for males. Marshall and Alcock (1981) found that males' would perform a "land and walk" response at their focal point significantly more often when presented dummy yellow models than to models painted black. Males' detected the models within 1-2 meters of the focal point. Similar results have been obtained with live and freshly killed tethered bees instead of models (Buchmann and Minckley, unpubl.), It thus appears visual signalling is effective for only short distances and cooration may operate primarily as close range location by incoming females. Visual signals by themselves do not appear to be sufficient to attract females from distances greater than several meters and since females are not observed hovering outside male focal areas it would appear another cue is used for long distance signalling. Visual signalling is also not consistent with the male display behaviors and location leks are positioned in plants. The flying and hovering behavior 74 15- 13 IO- I 3- I 2. 3 4 NLJM&EPi OP £>ITE€> Figure 16 Frequency of the number of sites used per day. 75 (/> CD O 14 B £ A * Oleander 12 B * Ornamental pistachio co> C 3 Mesquite •>» 10 D » Fig 0 E * Chinaberry Q. 8 CO F «Palo verde •a 6 M— O 4 a> 2 JD E 0 n J=L a aE1Q nnn EL A B C 0 E F A| A2 B| BJ> C| CJ CJ C4 D| OG E F Z Site Station Figure 17 Number of days sites were used in 1984 mating season. 76 MAL-E USING- PP»EPEPiP»ED &ITE US»IN<5r ANV eiTE IO 5 Z. 4- NUME?0.P-) OF MALE€? Figure 18 Frequency of numbers of males displaying per day compared with days the preferred site was also used, excluding the chinaberry site. 77 is highly energetic and 25% of the time a male is stationary alongside his focal point. Furthermore, focal areas and corresponding focal points were consistently located beneath the crest of the tree and if visual signalling was a primary concern to males these sites would be expected to be at tree tops or other areas unimpeded by adjacent vegetation. This, however, does not appear to be happening. The recent discovery of a large dorsal mesosomal gland within the dorsal thoraces of male X. varlpuncta appears to provide odoriferous material for such long-range chemical signals. Although appropriate bioassay experiments are necessary to confirm this function, we believe that the 3-component exocrine product blend is a sex attractant (Andersen et. al., 1988). Field evidence strongly supports this proposition. Marshall and Alcock (1981) were the first to describe a "strong flowery scent downwind of displaying males" and I have reliably located dozens of territorial males by the same odor which is detected on undissected males and those that have had glands opened. The odor plumes can be detected more than 12 m downwind. Besides producing and storing a large amount of secretion (perhaps 20 to 40 microliters, S. Buchmann unpubl.), the gland opens within a sulcus located on the mesosomal thoracic segment posterior to the wing bases. With wings serving as fans, male display may be viewed as a behavioral method to disperse the secretion. For mate encounters, a need for a long distance signal is evident in this species. Females are strong fliers and disperse widely. The central theme of male mating behavior is display behaviors which disseminate a pheromonal and visual signal to potential mates. 78 This is consistent with male-male activity within territories. Time budget data showed that males spent most of their time on display behaviors and only 16% of the time were engaged in territorial defense. Therefore, in this system male-male competition isnot concerned with direct physical aggression. Time and energy spent on display behaviors as well as male morphology implicate that males are competing for female attraction via various mechanisms of male advertisement. Despite this, there was aggression at territories and males compete for these sites although resident males usually maintained territory control (Table 8). Indications were that resident males preferred no other males to be closeby. Only rarely did I observe males to display in close proximity. On days that several males did display near one another, the owner of the preferred territory would leave his focal area and pursue any other male using one or both types of territorial aggressivity. This indicates males compete for space beyond their focal area. Most male-male behaviors are clearly for purposes of territory usurpation or defense. This is true even for the aggressive behavior I described which was curiously nonaggressive and long lasting (3-5 minutes). I consider these to be "contests" since one male always departed. Also, only once did I observe more than one male hovering in the same bush after these interactions, as often did occur after the escalated fights, suggestive that they are aggressive encounters. The slgnifigance of the large number of low level agonistic territory disputes (Table 8) may be that territory value is low and furthermore the males' low probability in obtaining matings. Escalated defensive or 79 offensive behaviors may be unreasonably costly since this would not directly increase mating success. Certainly, no observations have indicated that females view "battles" or choose amongst males based on encounter outcomes. Decreased territory value is proposed by Alcock and Smith (1986) to explain the increased numbers of multiple males displaying in single bushes they observed towards the end of the mating season. Male mating behavior does not by itself give an accurate indication of what males compete among themselves for. It is clear that display behaviors take up most of their time and it appears that territorial defense has been economized since oftentimes neither the attack nor the defense is very vigorous. These data suggest that a premium is placed on male advertisement and less value on territories. Males, however, do contest territory ownership indicating they mutually recognize a common characteristic(s). These may be some attribute of the site, of the territory or still be physical dominance display to females. This last option, however, is not probable as discussed above. Site Usage Patterns and Territorial Placement Tactics Despite the short length of the transect and small number of sites available for males to display, this site was useful for examining site selection in X. varipuncta because the site was flat and nest locations were known. Male display site location for this species has been discussed in three other studies (Alcock and Smith, 1986; Marshall and Alcock, 1983; Minckley and Buchmann, In prep) but all were in elevationally varying terrain and only Minckley and Buchmann (In prep) 80 knew the nest location. A comparison of findings between all studies should provide information toward an understanding of why particular sites are chosen. Particularly testable is Alcock and Smith's (1986) theory that hilltop or ridge topography acts to funnel females towards male territories. Seasonal site usage showed that male preference for sites did occur. Although in each year a different site was the most often used, it is clear that a single plant (site) and a particular position in the plant (station) was repeatedly used more than other sites and stations in this study area. These sites, which I consider to be preferred locations, were used on 64% of the total days any male hovered and on 70% of the days a single male was observed displaying in the study area. It is not known whether the same or different males are reusing sites day after day or if new males sometimes occupy these sites. To determine this, individual marked bees would have been necessary, which was tried but terminated because males' departed for the remainder of the afternoon. There is inferential evidence that the same individuals repeatedly used sites each day from the station data. Posts which occupy one position at a single plant, if used more than once, tended to be used over a sequence of days rather than different stations used on alternating days, or some other random pattern (Fig. 14). Site use was heirarchical since males were attracted to primary sites and established territories peripherally to these. This was first evidenced by data on the preferred territory which was used 9 of the 10 days I observed more than a single male displaying in the study area suggesting other males were attracted to this site. Secondary 81 territories are also more transitory and account for most of the territories used for less than 40 min. in Figure 15 suggesting that their territory is devalued relative to that of the preferred territory. The presence of these peripheral territories also indicates that an alternative mating tactic may occur in X. varipuncta. A subpopulation of males may continually sample potential territories and evaluate sites early in the season, afterwards returning to preferred sites and display either in these or, if repelled by the previous owner, display nearby. However, sneaky males are not likely in this system since females control mating so completely and should refuse nondisplaying males. The preferred male mating tactic of X. varipuncta is one of individual display. Daily, only a single male displayed in the study area on half the days (Fig. 13). This suggests that males are spacing themselves in the environment and agrees with Marshall and Alcock's (1983) assessment that male X. varipuncta use a dispersed lek territorial strategy. Clumped distribution of territories have not been found in the vicinity of nests (Minckley and Buchmann, in prep; this study) or in washes (Marshall and Alcock, 1981). On 10 days over the two year period multiple males were present. The maximum number seen in the area behind the houses was 4 males on one day and 5 males were seen on this same day if the chinaberry trees containing the nests are included also. Questions still remain concerning what mechanism males use to attract females to their territories. Territory placement in the environment is often used as an indication of what these mechanisms are. Males should locate themselves so to optimize the effectiveness of any 82 mechanism so that it attracts the maximum number of females (Thornhill and Alcock, 1983). For X. varipuncta, work by Alcock and Smith (1986) has addressed this topic in the same manner as have theoretical discussions (Emlen and Oring, 1977; Parker, 1978; Bradbury and Gibson, 1985; Gibson, Bradbury and Liu, 1986) that male distribution should reflect the distribution of females in the environment. However, my data are not consistent with a female distribution theory of male lek site positioning. If the female distribution theory was entirely correct, then X. varipuncta male territorial sites would be predicted to coincide with nest aggregations, which they do not. This is logical because X. varipuncta females are highly polylectic (Hurd and Linsley-, 1975) and do not aggregate at specific plant resources. Furthermore, many nests are often found in a single tree or area (pers. obs.), probably due to a philopatric tendencies of females, and these nest aggregations do result in cohabitation of many females. Therefore the highest consistent concentration of females is along pathways leading to and from nest sites or at nest themselves. It is here that males would be expected to have their defended territories. However this is not the case. There is no evidence of nest territoriality in my observations or from other published reports. All studies have instead shown a signifigantly higher usage of tall trees, ridgetops and hilltops. The two studies in which nest locations were unknown found male densities to be much higher along ridgetops (Alcock and Smith, 1986) and in washes with tall ironwoods (Alcock and Smith, 1986; Marshall and Alcock, 1983). When nest location was known, hilltops and tall plants again were more desirable for males 83 than apparently suitable plants closer to the nest (Hinckley and Buchmann, in prep). In this study, X. varipuncta territories around nest sites were not clumped and males did not appear to preferentially display along activity paths. Thus territory placement is not concordant with theoretical predictions that males establish territories in areas of high female density. Male spacing was enforced through male-male competition and resulted in a single male displaying around the nest each day. The criteria for territory site placement appears to be related to transmission of their sex pheromone signal. All authors have noted males display in tall plants or along ridgetops (Alcock and Smith, 1986; Marshall and Alcock, 1983; Minckley and Buchmann, in prep.). Products from the dorsal mesosomal gland are consistent with behavioral studies made before and after the glands' discovery that males disseminate a strong flowery aroma which probably acts as a female attractant. Both ridgetops and tall trees are relatively higher than their surroundings and serve to facilitate the projection of pheromones into the air column. It is this feature, which male display sites in all the studies had in common. Presence of the gland may also explain the type of display behaviors observed, since wings may aid in dispersing the chemical from the body. If aerial transmission of a pheromone is the key to understanding male X. varipuncta territorial behavior and lek site placement in the environment, as I argue, this suggests that female distribution is not a primary criteria. Although it could be argued that female distribution is the same and thus optimal over a wide area once away from the nest 84 (because females ability to disperse for long distances) this does not explain why territories are not situated more tightly near nests. Dispersed territories, aerial height in tall plants or in plants along ridgetops and the preponderance of time spent in display behaviors all indicate males compete for and are most concerned with those locations where pheromonal signals can be effectively transmitted to as many females as possible. Once away from nests, many sites may be optimally situated according to female distribution and male competiton may be for taller plants or in variable terrain, plants along ridgetops. Based on territory dispersion and gland size, pheromone plumes emanating from a single male appear to be adequate to communicate his location to any female which may encounter the extensive chemical plume. Female distributions may then be highly influenced by pheromonal signals and not by the topographic features as postulated by Alcock. and Smith (1986). LITERATURE CITED Alcock, J. 1987. Leks and hilltopping in insects. J. Natural History 21:319-328 Alcock, J., E.M. Barrows, G. Gordh, L.J. Hubbard, L. Kirkendall, D.V. Pyle, T.L. Ponder and F.G. Zalom. 1978. The ecology and evolution of male reproductive behaviour in the bees and wasps. Zool. J. Linnaean Soc. 64:293-326 Alcock, J. and A.P. Smith. In press. Hilltopping, leks and female choice in the carpenter bee, Xylocopa varipuncta (Hymenoptera: Anthophoridae). J. Zool. Altmann, J. 1973. Observational study of behavior: sampling methods. Behavior 32:223-267 Andersen, J.F., S.L. Buchmann, D. Veisleder, R.D. Plattner and R.L. Minckley. 1988. Identification of thoracic gland constituents from male Xylocopa spp. Latreille (Hymenoptera: Anthophoridae) from Arizona. J. Chem. Ecol. Anzenberger, G. 1977. Ethological study of African carpenter bees of the genus Xylocopa (Hymenoptera, Anthophoridae). Z. Tierpsychol. 44:337-374 Baird, J.M. 1986. A field study of thermoregulation in the carpenter bee, Xylocopa virginica virginica (Hymenoptera: Anthophoridae). Physiol. Zool. 59(2):157-168 Barrows, E.M. 1980a. Results of the survey of damage caused by the carpenter bee Xylocopa virginica (Hymenoptera: Anthophoridae). Proc. Entomol. Soc. Wash., 82:44-47 Barrows, E.M. 1980b. Robbing of exotic plants by introduced carpenter and honey bees in Hawaii, with comparative notes. Biotropica 12(1):23-29 Bradbury, J.W. 1985. Contrasts between insects and vertebrates in the evolution of male display, female choice, and lek mating. In, Experimental Behavioral Ecology, eds. B. Holldobler and M. Lindauer, G. Fischer Verlag, Stuttgart, New York Bradbury, J.W. and R.M. Gibson. 1985. Leks and mate choice. In, Mate Choice, ed. P. Bateson. Cambridge Univ. Press, Cambridge 85 86 Bradbury, J.V. and S.L. Vehrencamp. 1977. Social organization and foraging in Emballonurid bats. III. Mating systems. Behav. Ecol. Sociobiol. 2:1-17 Cane, J.H. 1981. Dufour's gland secretion in the cell lining of bees (Hymenoptera: Apoidea). J. Chem. Ecology. 7:403-410 Chappel, M.A. 1982. Temperature regulation of carpenter bees (Xylocopa californica) foraging in the Colorado desert of southern California. Physiol. Zool. 55(3):267-280 Cruden, R.W. 1966. Observations on the behavior of Xylocopa c. californica and X. tabaniformis orpifex (Hymenoptera: Apoidea). Pan-Pac. Entomol. 42:111-119 Daly, H.V., C.D. Michener, J.S. Moure and S.F. Sakagami. 1987. The relictual genus Manuelia and its relation to other xylocopinae (Hymenoptera: Apoidea). Pan-Pac. Ent. 63(2): 102-124 Emlen, S.T. and L.W. Oring. 1977. Ecology, sexual selection, and the evolution of mating systems. Science 197: 215-223 Fowler, H.G. 1983. La abeja Xylocopa californica arizonensis (Hymenoptera: Anthophoridae) e Yucca elata (Agavaceae); nidos, poblaciones, comportamiento e importancia al ciclaje de nutrientes en el desierto de Chihuahua. Folia Entomologica Mexicana. 56:75-83 Gentry, H.S. 1982. Agaves of Continental North America. Univ. of Az. Press. Tucson, Az. 1-670 Gerling, D. 1982. Nesting biology and flower relationships of Xylocopa sonorina Smith in Hawaii (Hymenoptera: Anthophoridae). Pan-Pac. Entomol. 58(4):336-351 Gerling, D. and H.R. Hermann. 1978. Biology and mating behavior of Xylocopa virginica L. (Hymenoptera: Anthophoridae). Behav. Ecol. Sociobiol. 3:99-111 Gerling, D., P.D. Hurd, Jr. and A. Hefetz. 1981. In-nest behavior of the carpenter bee, Xylocopa pubescens Spinola (Hymenoptera: Anthophoridae) J. Kans. Ent. Soc. 54(2): 209-218 Gerling, D., P.D. Hurd, Jr. and A. Hefetz. 1983. Comparative behavioral biology of two Middle East species of carpenter bees (Xylocopa Latreille) (Hymenoptera: Apoidea). Smithsonian Contr. Zool. No. 369. Smithsonian Institute Press, Washington D.C. 87 Hefetz, A., M.S. Blum, G.S. Eickwort, J.W. Wheeler. 1978. Chemistry of the Dufour's gland secretion of halictine bees. Comp. Biochem. Physiol. 61B:129-132 Heinrich, B. and S.L. Buchmann. 1986. Thermoregulatory physiology of the carpenter bee, Xylocopa varipuncta. J. Comp. Physiol. B. 156:557- 562 Howell, D.J. and B.S. Roth. 1981. Sexual reproduction in agaves: the benefits of bats; the costs of semelparous advertising. Ecology 62:1-7 Hurd, P.D., Jr. 1955. The carpenter bees of California. Bull. Calif. Insect Survey 4:35-72 Hurd, P.D., Jr. 1958. Observations on the nesting habits of some New World carpenter bees with remarks on their importance in the problem of species formation (Hymenoptera: Apoidea). Ann. Ent. Soc. Amer., 51:365-375 Hurd, P.D., Jr. 1959. Beefly parasitism of the American carpenter bees of the genus Xylocopa Latreille (Diptera; Hymenoptera). J. Kans. Ent. Soc., 32:53-58 Hurd, P.D., Jr. 1978. An Annotated Catalog of the Carpenter Bees (Genus Xylocopa Latreille) of the Western Hemisphere (Hymenoptera: Anthophoridae), Smithsonian Institute Press, 1-112 Hurd, P.D., Jr. and E.G. Linsley, 1975. The principle Larrea bees of the southwestern United States (Hymenoptera: Anthophoridae). Smithsonian Contr. Zool. 193:1-74 Hurd, P.D., Jr. and J.M. Moure. 1963. A Catalog of the Large Carpenter Bees belonging to the Genus Xylocopa Latreille (Hymenoptera: Apoidea). Univ. Calif. Publ. Entomol. 29:1-365 Iwata, K. 1964. Egg gigantism in subsocial hymenoptera, with ethological discussion on tropical bamboo carpenter bees. Nature and Life in Southeast Asia (Kyoto), 3:399-434 Janzen, D.H. 1966. Notes on the behavior of the carpenter bee, Xylocopa fimbriata in Mexico (Hymenoptera: Apoidea). Kans. Ent. Soc. 39:633-640 Krombein, K.V. 1967. Trap Nesting Bees and Wasps: Life Histories, Nests and Associates. Washington D.C.: Smithsionian Institute Press Kronenberg, S. and A. Hefetz. 1984. Comparaqtive analysis of the Dufour's gland secretion of two carpenter bees (Xylocopinae: Anthophoridae) with different nesting habits. Comp. Biochem. Physiol. 79B:421-425 88 Levene, H. 1953. Genetic equilibrium when more than one ecological niche is available. Am. Nat. 87:331-333 Levin, G. 1968. Evolution in Changing Environments. Princeton Univ. Press, Princeton, N.J. Linsley, E.G. 1965. Notes on the male territorial behavior of the Galapagos carpenter bee (Hymenoptera: Apidae). Pan-Pac. Entomol. 41:158-161 Louv, G.N. and S.W. Nicholson. 1983. Thermal, energetic and nutritional considerations in the foraging and reproduction of the carpenter bee, Xylocopa capitata. J. Ent. Soc. Sth. Afr. 46(2):227-240 Mackay, W.P., S. Silva, D.G. Lightfoot, M.I. Pagani and W.G. Whitford. 1986. Effect of increased soil moisture and reduced soil temperature on a desert soil arthropod community. Amer. Midland Nat. 116(l):45-56 Marshall, L.D. and J. Alcock. 1981. The evolution of the mating system of the carpenter bee Xylocopa varipuncta (Hymenoptera: Anthophoridae). J. Zool., London 193:315-324 Marston, N. 1970.Revision of the New World species of Anthrax (Diptera: Bombyliidae) other than the Anthrax albofasciatus group. Smithsonian Contrib. Zool. no. 43 May, D.G.K. 1974. An investigation into the chemical nature and origin of the waxy cell lining of the brood cells of a sweat bee, Augochlora pura (Hymenoptera: Halictidae). J. Kans. Ent. Soc. 47:504-516 Maynard Smith, J. 1966. Sympatric speciation. Amer. Nat. 100:637-650 Michener, C.D. 1972. Direct food transfering behavior in bees. J. Kans. Ent. Soc. 45:573-576 Michener, C.D. 1974. The Social Behavior of Bees. Harvard Univ. Press. Cambridge, Mass. Michener, C.D. 1979. Biogeography of the bees. Ann. Missouri Bot. Gard., 66: 277-347 Michener, C.D. 1985. From solitary to eusocial: need there be a series of intervening species. In, Experimental Behavioral Ecology, eds. B. Holldobler and M. Lindauer. G. Fischer Verlag, Stuttgart, New York Michener, C.D., R.B. Lange, J.J. Bigarella and R. Salamuni. 1958. Factors influencing the distribution of bees' nests in earthen banks. Ecology 39(2):208-217 89 Minckley, R.L. New host record and biological notes of the Bombyliid flies, Anthrax simson habrosus (Marston) and A. xylocopae (Marston) (Diptera: Bombyliidae), (In prep.) Minckley, R.L. and S.L. Buchmann. Lek site selection of male Xylocopa varipuncta (Hymenoptera: Anthophoridae), (In prep). Morse, D.H. 1980. Behavioral Mechanisms in Ecology. Harvard Univ. Press. Cambridge, Mass. Nicholson, S.V. and G.N. Louve. 1982. Simultaneous measurement of evaporative water loss, oxygen consumption and thoracic temperature during flight in a carpenter bee. J. Exp. Zool. 222:287-296 O'Neal, R.J. and G.D. Waller. 1984. On the pollen harvest by the honey bee (Apis mellifera L.) near Tucson, Arizona (1976-1981). Desert Plants 6(2):81-110 Parker. G.A. 1978. Evolution of competitive mate searching. Ann. Review of Ent. 23:173-196 Pepper, J.H. and E.Hastings. 1952. The effects of solar radiation on grasshopper temperatures and activities. Ecology 33:96-103 Rau, P. 1933. The jungle bees and wasps of Barro Colorado Island (with notes on other insects). Kirkwood, St. Louis Mo.: Published by P. Rau. pp. 1-324 Rubenstein, D.I. and R.W. Wrangham. 1986. Ecological aspects of social evolution. In, Social Evolution in Birds and Mammals, eds, D.I. Rubenstein and R.W. Wrangham. Princeton Univ. Press. Princeton, N.J. Schaffer, W.M. and M.V. Schaffer. 1979. The adaptive signifigance of variations in the reproductive habit in Agavaceae II. Pollinator foraging behavior and selection for increased reproductive expenditure. Ecology 60:976-987 Seely, T.D. and P.K. Visscher. 1985. Survival of honeybees in cold climates: the critical timing of colony growth and reproduction. Ecol. Entomol. 10:81-88 Seger, J. 1985. Intraspecific resource competition as a cause of sympatric speciation. Evolution-essays in honour of John Maynard Smith. University Press. Cambridge. Skaife, S.H. 1952. The yellow-banded carpenter bee, Mesotrichia caffra Linn, and its symbiotic mite, Dinogamasus braunsi Vitzthun. J. ent. Soc. sth. Afr. 15(1):63-76 90 Smith, W.E. and W.G. Whitford. 1978. Factors affecting the nesting success of the large carpenter bee, Xylocopa californica arizonensls. Environ. Entomol. 7(4):614-616 Vluegal, D.A. 1947. Waaremingen aan het gedrag van de Grijze Graafbig (Andrena vaga Panz.) (Hym.) Entomol. Ber. 12:185-192 Velthius, "H.H.W. and J.M.F. de Camargo. 1975. Further observations on the function of male territories in the carpenter bee Xylocopa (Neoxylocopa) hirutissima Maidl (Anthophoridae, Hymenoptera). Netherlands J. Zool. 25(4)s516-528 Velthius, H.H.W., Y. Wolf and D. Gerling. 1984. Provisioning and preparation of the brood cell in two carpenter bees, Xylocopa sulcatipes Maa and Xylocopa pubescens Spinola (Hymenoptera: Anthophoridae). Israel J. of Ent. 18:39-51 Vinson, S.B., G.W. Frankie and H.J. Williams. 1986. Description of a new dorsal mesosomal gland in two Xylocopa species (Hymenoptera: Anthophoridae) fom Costa Rica. J. Kans. Ent. Soc. 59:185-189 Waring, G.L. and R.L. Smith. 1986. Patterns of faunal succession in Agave palmeri. Southwestern Nat.