|UFDC Home||myUFDC Home | Help ||
|Table of Contents|
|Numbers of Migrants|
|Phenology of Movement|
|Summary and Discussion|
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
STANDARD VIEW MARC VIEW
|Table of Contents|
Table of Contents
Numbers of Migrants
Phenology of Movement
Summary and Discussion
BUTTERFLY MIGRATION THROUGH THE FLORIDA PENINSULA
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
This work is lovingly dedicated to my parents, Lucjan and Stanislawa Lenczewski, who have waited half their lives to see it and to my husband, Joseph Jowers, for believing that he would.
A number of people were helpful during this research. Foremost, I would like to thank my major professor, Dr. Thomas J. Walker, for supplying the materials and workshop used in construction of the portable traps and for providing expertise and labor during their construction. The funds for travel and other expenses were arranged by Dr. Walker and he provided a great deal of enthusiasm and encouragement for this project throughout many years. I am also indebted to Dr. J. J. H. Frank, who, through the Mole Cricket Biocontrol Project, provided the opportunity for a graduate assistantship and facilitated the loan of travel vehicles. Dr. Lincoln P. Brower provided many opportunities for "butterfly" socializing, and the use of his property for mark-release-recapture studies, during which he kindly made many valuable observations. Dr. Brower and Dr. Frank Slansky generously allowed me the use of their laboratories and equipment. I also thank the other members of my committee, Drs. J. E. Lloyd and J. L. Nation, for their helpful suggestions and a critical review of the manuscript.
The following people made the trapping study feasible by providing locations for, and assistance with, flight traps. I am particularly grateful to Dr. J. Whitesell, Valdosta State College; Dr. G. W. Ellstrom and his assistant, Ms. Annette Chandler, JFAS/AREC Leesburg; Dr. Mohammed Abou-Setta and Dr. H. N. Nigg, IFAS/CREC; and Dr. Mark Deyrup, Archbold Biological Station. Special thanks go to Dr. Whitesell
for field assistance with traps, Dr. Nigg for a tour of the Green Swamp and Dr. Deyrup for help during hurricanes and otherwise.
My first interest in butterflies and their migration through the Florida Peninsula was inspired by the late Dr. Dennis Leston, in whose memory this work is completed. His enthusiasm for nature and love for insects will remain with me always. I am greatly indebted to my parents, Lucjan and Stanislawa Lenczewski, who have encouraged my education. My sister, Eva Lenczewski, despite her dislike of insects, has sometimes even helped with this perversity of mine and she deserves recognition for her courage. Finally, and most importantly, I thank my husband, Joseph Jowers, for a great deal of assistance and companionship in the field. He and my son, Justin, held down the home front to really make my work possible.
TABLE OF CONTENTS
CHAPTER 1. INTRODUCTION.................................. 1
Literature Review ....................................... 1
General Migration................................... 1
Florida Butterfly Migration ............................. S
Research Plan.......................................... 10
Biology of the Butterflies................................... 11
Phoebis sennae..................................... 11
Agraulis vanillae.................................... 15
yjrJarms pipjejis.................................... 17
Precis penia...................................... 18
Other Migrants..................................... 20
CHAPTER 2. FLIGHT DIRECTION ...............................23
Materials and Methods ....................................25
Results and Discussion ....................................26
CHAPTER 3. NUMBERS OF MIGRANTS ...........................38
Materials and Methods ....................................39
Latitudinal Pole Counts................................39
Results and Discussion....................................44
Latitudinal Pole Counts................................44
Cloudless sulphur............................... 44
Gulf fritillary ................................. 46
Long tailed skipper.............................. 46
Migration Profiles................................... 49
Cloudless sulphur............................... 49
Gulf fritillary ................................. 58
Numbers of Migrants................................. 66
CHAPTER 4. PHENOLOGY OF MOVEMENT ........................77
Materials and Methods ....................................80
Longitudinal Pole Counts ..............................80
Portable Flight Traps.................................81
Results and Discussion ....................................83
Longitudinal pole counts ..........................83
Flight trap catches ..............................85
Gulf Fritillary .....................................93
Longtiudinal pole counts ..........................93
Flight trap catches ..............................96
Long Tailed Skipper ................................102
Longitudinal pole counts ......................... 102
Flight trap catches .............................105
Flight trap catches .............................112
CHAPTER 5. SUMMARY AND DISCUSSION........................ 118
Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy
BUTTERFLY MIGRATION THROUGH THE FLORIDA PENINSULA
Barbara Lenczewski December 1992
Chairman: Dr. Thomas J. Walker
Major Department: Entomology and Nematology
At least eight species of butterflies migrate southward during fall of each year
through the Florida peninsula. The four major migrants, representing four families of
Lepidoptera, are Phoebis sennae (L.) (Pieridae), the cloudless sulphur; Agraulis vanillae
(L.) (Heliconiidae), the gulf fritillary; Urbanus proteus (L.) (Hesperiidae), the long tailed
skipper; and Precis coenia (Hubner) (Nymphalidae), the buckeye. Drive counts of the
net southward movement of the cloudless sulphur and gulf fritillary were made across
north central Florida at the latitude of Gainesville (29.65) during fall 1986-1988. This
migration profile, or cross section through the migratory stream, was used to estimate
the density of individuals moving south into central Florida annually. The highest
density of the cloudless sulphur was through the central parts of the state and that of the
gulf fritillary, along the western sections. Peak numbers were recorded just east and just
west of Gainesville, respectively. Both species avoided the Atlantic coast. Using previous data from Gainesville flight trap catches in conjunction with density estimates from this study, the total number of cloudless sulphurs and gulf iritillaries moving south each fall into central Florida was estimated at 42 and 115 million individuals, respectively. When previous estimates for the long tailed skipper and buckeye were included, the total number of butterflies migrating south through the Florida peninsula was estimated to be 175 million individuals annually. Observations of flight azimuths for three species (excluding the buckeye) were made at various locations south of Gainesville during fall 1986. Most significant mean flight directions were consistent with the 141 previously described for Gainesville and the limit of significant southward movement for that year was observed just south of Lake Ockeechobee for all three species. The phenology of the migration was studied by trapping the four major migrant species using directional flight traps at five stations from south Georgia to south central Florida during fall 1987-1988. All significant net movement was southward, extending to the Lake Alfred site, about 192 km south of Gainesville. There was no evidence from trap catches or drive counts at the Lake Placid latitude (27.45) that large numbers of migrants move into extreme south Florida to overwinter.
CHAPTER 1 INTRODUCTION
Migration has been recognized in recent years as an important element in the dynamics of insect populations (Solbreck, 1985). Long distance movements of butterflies (Williams, 1930), locusts (Johnson, 1969), aphids (Kennedy, 1961), moths (Stinner si al, 1983), dragonflies (Dumont and Hinnekint, 1973) and insects of many other groups have been recorded by observers across practically every continent (Williams, 1958), at high altitudes (Rainey, 1951), and many miles out to sea (Baust el al, 1981). Although some insect flights seem to function as random dispersal by forces beyond the individual's control (Johnson, 1969), closer behavioral studies reveal that active initiation of flight occurs in many species and correlates with favorable wind conditions and appropriate seasons (Baker, 1978; Taylor and Reling, 1986). In other words, the migrants will "choose" the appropriate winds, during the appropriate times of year, by which they may be carried.
It was perhaps due to his interest in butterflies that Williams (1930) stressed the concept of individual control over flight direction in his early definition of migration. This individual control over flight direction is particularly evident in butterfly migrations.
These highly visible insects travel within the boundary layer, the layer of air near the
ground where wind velocity is less than the insect's air speed, and flight direction can
be controlled by the individual. The thickness of this boundary layer is variable,
determined by wind velocity and the air speed maintained by the individual (Taylor,
1958; Pedgley, 1982). This determined choice of flight track is perhaps easiest to
observe in butterflies and demonstrates that migration can be more under the individual
control of these, and perhaps other, insects than previously suspected.
There has been considerable controversy in theoretically extricating dispersion
from migration. Baker (1978) has suggested that there is a continuum in the expression
of the migratory habit, ranging from random movement to the evolution of highly
specific destinations with correspondingly adaptive physiological changes. In a recent
summary, Dan than arayana (1986, p. 1) claimed a consensus has emerged on the
terminology used to describe insect movements. His definition of migration will be
accepted for the purposes of this study:
Non-migratory movements involve travel within the habitat associated with such activities as feeding, mating and oviposition ... In contrast, migratory movements take insects beyond the habitat for the purposes of colonizing new habitats, re-colonizing old ones, aestivation or hibernation ....
The defining criterion is travel outside what is, or had previously served, as the
sustaining "habitat" to more favorable conditions. The behavior and physiology involved
in getting to these new habitats fall somewhere on a continuum in the evolution of the
migratory habit. Some insect migrants, such as the monarch butterfly, may be
reproductively inactive during these movements and maintain long distance flights on
stored body fats, but in others, mating and feeding occur in conjunction with the
migratory movement. Although not necessarily so, a cyclic, or seasonal pattern is often seen in these migratory movements. The same individual may return, at least part way, to re-colonize old habitats, or, the return may consist of a step-wise expansion of new generations into the former breeding range during favorable conditions (Walker, 1985a).
Baker (1969) has speculated on the evolution of migration and the maintenance of migratory behaviors in populations. The factors that initiate migration are varied and have been found to be finely attuned by genetic programming (Lamb and McKay, 1983) to changes in habitat quality over time and space (Southwood, 1962) and other environmental cues (Dingle, 1972). The influence of genetic factors on migratory tendency in members of a population has been investigated empirically by Dingle t al (1977) and Istock (1978), and theoretically by Roff (1975). It is known that individual dispersal tendencies can be highly variable within populations (Baker, 1978) and these characters have been influenced by experimental selection to some degree (Dingle, 1968; Rankin, 1978). There are a number of possible life history strategies available to an insect and how natural selection operates on these genetic factors in the evolution of the migratory habit is still unclear.
How migrants find their destinations, that is, their navigational techniques, (Schmidt-Koenig, 1975), also remain a mystery. Monarch butterflies are highly specific in their annual migration to overwintering sites in Mexico (Calvert and Brower, 1980). Walker (1985a) suggested that cloudless sulphurs from the southeast navigate toward the Florida peninsula.
Migrating butterflies can maintain their compass direction and therefore must have some orientation mechanism. It is possible that a time-compensated sun compass is used (Walker, 1985a; Baker, 1978), but although Baker (1968a) has shown possible azimuth orientation to the sun's position, by Pieris rapae. this remains to be proven (Able, 1980). That fall migrants can maintain their migratory direction on overcast days, although not many fly then, weakens, but does not eliminate, this theory (Verheijen, 1978). Another possibility is that they have a magnetic compass. Evidence of geomagnetic orientation has been found in the underwing moth, Noctua pronuba L. (Baker and Mather, 1982), and is well known in honey bees (Lindauer and Martin, 1972). Magnetic particles have been detected in the monarch butterfly, although their function has not been demonstrated (MacFadden and Jones, 1985). Jungreis (1987) tested two of the Florida migrant butterflies, Phoebis sennae (L.) and Agraulis vanillae (L.), and found no evidence of magnetic particles.
Migration requires a high degree of coordination among behavior, morphology and metabolism to function as an adaptive life history strategy. Migrant insects must undergo physiological changes that can sometimes be used to identify their migratory tendency. One metabolic preparation for sustained flight that has been studied is energy storage. A primary source of energy is fat, or lipids (Beenakkers, 1965), and increased levels have been found in migrating locusts and monarch butterflies (Johnson, 1974). It is known that monarchs build up these lipid stores prior to fall migration and deplete them during long flights to Mexico (Beall, 1948; Brower, 1977). Whether they replenish, or slow, the depletion of these stores by nectar feeding enroute has still not
been determined (Meier and Fivizzani, 1980). Beall (1948) reported that monarchs captured in Ontario, before flight, were heavier than those caught after flight in Louisiana. Fat stores are conserved carefully by thermoregulation through the overwintering period and must remain sufficient to at least fuel the return journey to new feeding or breeding grounds (Masters, 1988).
There has been very little experimental investigation of migration in the laboratory. Most studies in this area have been done with insects in which flight activity and/or orientation can be easily measured, and usually in tethered flight (Dingle, 1985). So far, researchers have been unable to measure a butterfly's level of flight activity and preferred direction in the laboratory. Flight traps (Walker, 1985b; Walker and Lenczewski, 1989), that exploit the behavioral characteristic of migrants to go up and over obstacles, rather than around, can segregate these individuals in the field, but the problems in the laboratory remain.
Florida Butterfly Migration
In a long term study using permanent flight traps in Gainesville, Walker (1991) has identified at least eight species of butterflies that migrate southward each fall through the Florida peninsula. The four principal migrants represent four families of Lepidoptera: Phoebis sennae (L.) (Pieridae), the cloudless sulphur; Agraulis vanillae (L.) (Heliconiidae), the gulf fritillary; Urbanus proteus (L.) (Hesperiidae), the long-tailed skipper; and Precis coenia (Hubner) (Nymphalidae), the buckeye. These species comprised 85% of Walker's (1991) total trap catch and they also exhibited bidirectional
seasonal movement i.e. north in the spring and south in the fall. The total size of the fall migration of these species has been estimated using preliminary data from this study by Walker (1991) at about 80 million individuals. This is comparable to the estimates for monarchs at Mexican overwintering sites (Calvert el si, 1979). The most significant and noticeable movements through the Florida peninsula are exhibited by the cloudless sulphur and the gulf fritillary, which have been estimated by Walker (1991) to comprise nearly 80% of the fall migration, or 64 million individuals annually.
A number of other species also show less directed or irregular migratory movements (Walker, 1991). These include four pierids, Pieris rapae (L.), the cabbage white; Eurema daira (Godart), the dainty sulphur; Eurema lisa (Boisduval & LeConte), the little sulphur; Eurema nicippe (Cramer), the sleepy orange, as well as a nymphalid, Vanessa virginiensis (Drury), the painted lady. Two hesperiid skippers are known to migrate southward in the fall through north Florida, although there is very little known of their movements since they are too small to be retained by Walker's (1991) traps. They are Lerema accius (Smith), the clouded skipper and Panoquina ocola (Edwards), the ocola skipper. Walker (1991) also reported some individuals of the monarch butterfly, Danaus plexippus (L.), as passing through Gainesville. The majority of the monarch migration, however, is southwest toward Texas and Mexico (Brower, 1977).
The recurring spectacle of huge butterfly migrations along the Gulf coast (Urquhart and Urquhart, 1976) and through north central Florida (Walker, 1978, 1980, 1985ab, 1991) is locally well known, but is inadequately represented in the literature. The Florida Gulf coast is a renowned stopover along the major North American
migratory route for birds (National Geographic, 1979). It seems to function this way, in some respects, for migrant butterflies also. Large numbers of monarchs, gulf fritillaries and long tailed skippers are often seen feeding at Baccharis halimifolia along the Gulf coast, particularly in the St. Marks or Appalachicola area (pers. obs.) during the fall. Cloudless sulphurs are also present, but not in such great numbers as the other species and there are some indications (Walker, pers. comm.) that they avoid coastlines. Despite the intriguing presence of these large aggregations of migrant butterflies which are obvious to the most casual observer, the subsequent movements of these individuals, except for the monarch, remain largely unknown. These species often exhibit erratic flight near the coastline, thus making it very difficult to find a preferred pattern. There are reports of movements west, following the coastline (Urquhart and Urquhart, 1976), but I have seen migration in both directions along the coast and also south across the water, as well as north back to land. The migrants seen massed along the Gulf are spectacular, and understandably, have attracted the most attention. Away from the Gulf coast, most north Florida residents realize that the numbers of these species increase in the fall, but are generally unaware of their persistent southward flights. Once the butterflies pass into the Florida peninsula, we know virtually nothing of the migration beyond Gainesville (Walker, 1991).
Most migrants pass through north peninsular Florida during September through November and it has been presumed they join resident populations in subtropical south Florida. Although the migratory flights are so conspicuous through north Florida during fall, this does not seem to be the case in the southern parts of the state. Although the
migrant species seem more common in south Florida during this time (Lenczewski, 1980), there has only been one report of Agraulis vanillae and Urbanus proteus migration in Melbourne (Williams, 1958). In addition, there are no data reporting any large aggregations of these species in south Florida, which is lepidopterologically very well explored, especially during the fall and winter months (Lenczewski, 1980). Perhaps, as suggested by Walker (1985a), the densities become insignificant when the migrants settle into the large areas that are available.
The only spring migrant trapped at Gainesville in large numbers has been Precis coenia. the buckeye. Although there must be some method of northern recolonization, there is only minimal direct evidence of spring return migrations by the long tailed skipper, Urbanus proteus. Spring migrations of Phoebis sennae and Agraulis vanillae are also small and little noticed except for significant flight trap catches (Walker, 1991). The total northern movement during spring of the four species has been estimated using preliminary data from this study as eight million individuals (Walker, 1991).
As stated previously, of the insect migrants, selection for precision in destination is perhaps most evident in the monarch butterfly (Brower, 1977), where large scale movements to several precisely located overwintering sites are coordinated for millions of individuals. The monarch completes the southward journey in reproductive diapause and the same individuals return, at least part of the way, during spring to repopulate the northern range (Brower, 1962). For the migrant butterflies observed moving through north Florida, it is clearly evident that the butterflies are mating, ovipositing and stopping to feed at nectar sources enroute. This reproductive activity must result in subsequent
generations that will also presumably join the migration south and perhaps colonize northern areas in which they have had no experience, the following spring. As Baker (1978) has suggested, the behavior and ecology of these species may demonstrate a sequence in evolutionary development along the spectrum of the migratory habit and can provide new insights into these behaviors.
Florida provides an ideal geographical location for the study of butterfly migration. The peculiar configuration of the state results in a number of migratory species from the southeastern United States being funneled through the peninsula presumably to the southern latitudes where they join resident populations and overwinter. Although there are reports of occasional butterflies some distance at sea (Baust gt al, 1981), the difficulty of long oceanic trips (Brower, 1962) most likely precludes the majority of butterflies from flying across the Gulf of Mexico, and certainly, the Atlantic. The peninsula of Florida, then, acts as a trap with a concentration of individuals contained by extensive water masses on three sides. The butterfly migration through Florida perhaps ranks among the world's phenomenal insect movements in that it involves a large number of unrelated species from a wide area within a short period of time. Larsen (1988ab) has reported a similar mixed butterfly migration during spring and fall in southern India. The movement involves more than twenty species, numbering approximately four million spring migrants and as many as 100 million fall migrants. The width of the migratory front reported by Larsen (1988ab) was, at most, only 12 km in the spring. It may also be true that the majority of the Florida spring return migration does not pass through Gainesville.
At the very least, understanding why, how and where these butterflies are going presents a fascinating puzzle. As stated previously, outside of Gainesville, Florida, virtually nothing is known of the phenology or extent of the migratory movements. The goal of this study was therefore to contribute to general knowledge of the migration, primarily of Phoebis sennae and Agraulis vanillae. elsewhere in peninsular Florida.
The investigations in this study focused on the migratory patterns in peninsular Florida of the two principal migrant butterflies, Agraulis vanillae and Phoebis sennae. Some attention was also given Urbanus proteus and Precis coenia. The research was designed as a four part study, carried out during the falls of 1986 through 1988 and addressed the following questions:
1. Flight direction Is the mean flight direction of migrants throughout Florida the same
as in Gainesville?
2. Numbers of migrants Is the migration density across Florida along the Gainesville
latitude constant? How many migrants pass into central Florida annually?
3. Phenology of movement Does the peak migration at southern sites occur at
approximately the same time as it does in Gainesville? Where does the migration stop?
Estimates of migration density for the cloudless sulphur and gulf fritillary were made along the Gainesville latitude (29.65). From these data, a "migration profile" or cross section, through the stream of migration for each species was constructed. This
information was used, together with Walker's (1991) data from Gainesville, to estimate the size of the annual fall migration of these two species moving through north central Florida. The phenology of the migration was studied by trapping the four principal migrant species using directional flight traps at five stations from south Georgia to south central Florida. Counts of north and south flying Phoebis sennae and Agraulis vanillae also were made on other west-east transects along the Florida peninsula.
The individual chapters have been written as drafts of separate publications. This has unavoidably resulted in some repetition. The numerical assignment of weeks for calendar dates during which sampling was done with all methods is given in Table 1-1. The location of all sites used for pole counts and flight trap stations are shown in Fig. 1-1.
Biology of the Butterflies
Commonly known as the cloudless sulphur, this is a fast flying butterfly of temporary, disturbed habitats. It is found in the eastern and southern United States, rarely straying to Canada, through the West Indies and south to Argentina (Lenczewski, 1980). In both the temperate and tropical zones, the subspecies of sennae exhibit strong migratory habits (Howe, 1975) and C.B. Williams (1938) nicknamed this "The Travelling Butterfly." The subspecies eubule occurs commonly in the southeastern United States, straying as far north as Canada (Klots. 1951) during the summer and fall
Fig. 1. I-Location of pole count and flight trap sampling locations used during the fall of 1986, 1987 and 1988.
Table 1-1. Calendar dates of sampling and the corresponding assigned week number for all sampling methods used during 1986-1988.
1986 1987 1988
1 01-07 SEP AUG 30-05 SEP AUG 28-03 SEP
2 08-14 06-12 04-10
3 15-21 13-19 11-17
4 22-28 20-26 18-24
5 29-05 OCT 27-03 OCT 25-01 OCT
6 06-12 04-10 02-08
7 13-19 11-17 09-15
8 20-26 18-24 16-22
9 27-02 25-31 23-29
10 03-09 NOV 01-07 NOV 30-05 NOV
11 10-16 08-14 06-12
12 17-23 15-21 13-19
13 24-30 22-28 20-26
14 29-05 DEC 27-02 DEC
15 06-12 03-10
months. Generally, the species occurs in the Mississippi Valley to Central Illinois and along the Atlantic coastal plain to New Jersey (Calhoun et al, 1990). It is known to reproduce as far north as west central Illinois (Sedman and Hess, 1985) and Virginia (Clark and Clark, 1951). In Florida, this species has been recorded for every month of the year throughout the state, but individuals are most common during the late summer and fall in north central Florida (Lenczewski, 1980; D. Baggett, pers. comm.). This subspecies is also resident in the states adjacent to Mexico (Howe, 1975). During some years, individuals are frequent in eastern Kansas and Missouri where they may be seen flying in a southeast direction in late summer (Howe, 1975). Similar flights have been
recorded in Mississippi and Alabama (Mather and Mather, 1958; Lambremont, 1968), apparently heading southeast toward the Florida peninsula. Walker's (1985a) observation of systematic changes in P. sennae's flight direction at many locations in the southeast suggests that these migrants are navigating toward peninsular Florida. The cloudless sulphur does not exhibit the coastal build up of other species mentioned and, as Walker has suggested (pers. comm.), may detect the "proper" time to change its flight direction before hitting the Gulf or Atlantic coastlines from up to 60 miles away! Gaddy and Laurie (1983) note some consistent discrepancies in flight direction along the South Carolina coastline for Phoebis sennae. In August, September and early October of 1978-1980, they noted northeastern migrations along the immediate coast and, what seemed to be random movements, inland. As previously discussed, flights adjacent to coastlines can be erratic. Migrating butterflies often seem "confused" and fly in what appear to be inappropriate directions, more often than not following the coastline in either or both directions.
Adult males are yellow above and unmarked. Females are a deeper, more orange yellow, fringed with dark brown marginal spots. The adults are particularly attracted to red flowers, many of which flower in the fall and are an important source of nectar for migrating butterflies during that time (pers. obs.). Larvae are a pale yellowish green with a yellow lateral stripe along each side. They are most often found at the young shoots of their leguminous hostplants and sometimes web leaves together for shelter (Howe, 1975). The larvae of this butterfly feed on a wide range of leguminous hostplants. There are at least 50 species of hostplants recorded and adults
will opportunistically oviposit on plants according to availability. The preferred larval hostplants in Florida for this species include various species of Cassia, particularly obtusifolia (L.) (pers. obs.), and as Howe (1975) reported in Louisiana, the partridge pea, Chamaecrista cinerea (L.). Two generations reportedly occur in the northern part of the range and although Howe (1975) reports that breeding is continuous in the Gulf region, including Florida, there are no records of such activity during the winter months, at least in the latter area. During 1987, Calhoun ej al (1990) reported cloudless sulphur individuals sighted in Wisconsin and New York. Numbers of individuals seen that summer were unusually large and sightings extended the northern range significantly. The species was common in Mississippi, Illinois, Kentucky, Indiana, Ohio and West Virginia. He described temporary breeding colonies of Phoebis sennae in Ohio and West Virginia, where apparently, due to the unusually warm winter of 1986-1987, individuals overwintered and were available in early spring to utilize newly emerged hostplants. The larval hostplants of this species usually senesce in late fall and are absent through the winter months. Interestingly, these breeding populations in Ohio and West Virginia were located along river banks, which Calhoun (1990) considered to be primary corridors of dispersal.
This butterfly, also known as the gulf friullary, is commonly found in temporary, disturbed habitats such as roadsides and fields, usually near its larval foodplant, during the late summer and fall months in Florida. The species is widespread,
ranging throughout tropical America, from Argentina north to the southern United States (Howe, 1975) and throughout the West Indies (Riley, 1975). The eastern subspecies, nigrior. occurs from Florida, west to Louisiana, north to North Carolina and is also found on Bermuda. Although this species has been reported as far north as New York (Howe, 1975), it is not commonly found above southern Virginia (Clark and Clark, 1951) on the eastern coast. Unlike the cloudless sulphur, this species is not usually seen during winter months (December-February) in northern Florida. However, there was an exceptionally warm winter in Gainesville during 1986-87 and adults were seen flying in January (pers. obs.).
Like the other heliconiids, the larvae of Agraulis vanillae feed only on species of Passiflora. In Georgia and north central Florida, the caterpillars are commonly found on the passion flower, Passiflora incarnata L. (pers. obs.). The larva has a black body, with three pairs of dorso-lateral, orange-brown stripes. There are six rows of branching tubercles on the body and a pair of longer tubercles curving back on the head. The adult is distinctively colored with bright orange upper wings marked with black and the hindwing underside covered with metallic silver spots and they are particularly attracted to the flowers of Spanish needle, Bidens pilosa and lantana, Verbena spp. According to Klots (1951), there are three or more broods in the south. In Gainesville, Florida, the butterflies arrive and begin to oviposit as soon as plants are available in early spring, continuing to breed until the winter months, usually late November (pers. obs.).
Urbanus proteus. the long tailed skipper, is a minor pest of cultivated beans, known as the bean leaf roller in agricultural literature. This species is distributed from Argentina through the United States and West Indies. The mainland populations comprise the nominate subspecies, with all of the insular populations falling into a second subspecies (Howe, 1975). In the United States, JJ. proteus proteus is found from Connecticut, south to Florida and west to Texas, Arkansas, Arizona and California. The larvae are green and have a dark mid-dorsal line with yellow lateral lines and green stripes below these. The area between the stripes is dotted with black and yellow spots. The head is large, reddish brown, with two yellow spots between the ocelli and the mouthparts. The caterpillars come out to feed from leaf shelters that they construct on their leguminous hostplants, moving to a larger one at each instar. The major foodplants include such legumes as Bauhinia. Clitoria mariana. Desmodium. Phaseolus. Soja. Vigna. Wisteria and Pueraria lobata (Lenczewski, 1980). In the adult, the thorax, wing bases and hind wings are a conspicuous metallic green. There are four separate square spots on the central band of the forewing and a larger square spot under the end of the cell. The male has a costal fold, the antennal club is yellowish brown below and the range of wingspread is 38-50 mm (Howe, 1975). Adults are found flying in open areas, particularly fields and along roadsides in disturbed sites. Three generations are reported in Florida but only representatives of the late-summer brood arrive in Virginia where this species is considered a casual visitor (Clark and Clark, 1951). Although Howe (1975) claims it occurs throughout the year, there are no records during the summer months of
June and July in south Florida (Dade or Monroe Cos.). Instead, the recently introduced, nonmigratory species, JJ. dorantes. with very similar habits is common during this period (Lenczewski, 1980).
Commonly known as the buckeye, this butterfly is the only Florida migrant which demonstrates relatively large scale spring northward migration (Walker, 1991). The genus Precis is largely confined to the tropical regions of the world, only two species occur in North America. There has been considerable confusion concerning the systemaucs of this species, which can be quite variable in size and color. P. coenia is found from southern Canada, west to California, Arizona and south through tropical America and Cuba. Clark and Clark (1951) reported two seasonal forms in Virginia. The spring/autumn form had an average forewing length of 27 mm in females and 24 mm in males. The summer form averaged a slightly greater wing length and darker color, characterized by an irregular reddish band on the underside of the hind wings which the spring/autumn form lacks. Clark and Clark (1951) noted that the wings in nearly all of the spring/autumn specimens they collected in Virginia were damaged. These worn wings may indicate long distance travel and this corresponds well to what has been observed of migration periods for this species. Mather (1967) found light and dark forms to be correlated with seasonal changes. He found the major seasonal shift from dark to light occurs between March and April and from light to dark between August and September. The shifts were found to coincide with a change in the mean
temperature from below 16C to above 16C and from above 27C to below 27C and not with rainfall, as previously suspected. Scott (1975) investigated the issues of male territoriality and migratory tendency which seem to be at theoretical odds. To the casual observer, buckeye males seem highly territorial, dashing from their resting places, to chase intruders or meet females. Scott (1975) concluded that it is a "psuedo-territorial" tendency and is not maintained for any significant length of time at a particular site.
The adults are often seen in open country, especially sitting along dirt roads near disturbed, weedy fields. Three broods occur in the southeast (in Virginia from late May to late fall). Although the butterflies of the last brood may be seen on warm days through the first half of winter, they are not found in the spring in Virginia (Clark and Clark, 1951). This species is considered a summer breeding resident in Virginia and there are no records of overwintering individuals. However, Clark and Clark (1951) report having seen ragged, overwintering individuals in March, April and early May in Washington, D.C. Howe (1975) describes the adults as hibernating and sometimes migrating. In southern Florida at least, the subspecies, P. coenia coenia. is reported for every month of the year (Lenczewski, 1980).
The larva has a dark gray body, striped or spotted with orange-yellow. There are a number of short, branching spines on the body and one pair on top of the head. The larval foodplants are Ruellia (Acanthaceae); Plantago (Plantaginaceae); Antirrhinum. Buchnera. Gerardia harperi. Linaria. Mimulus. Scrophularia lanceolata and others (Scrophulariaceae); Ludwidgia (Onagraceae); Sedum (Crassulaceae); Verbena prostrata
and Acuba (Verbenaceae). The Scrophulariaceae appear to be the most important larval hostplants in Florida.
Pieris rapae (L.). An introduction from England into Quebec in 1860, this well-known migrant has spread rapidly throughout most of North America. Also known as the cabbage white, it can be a serious agricultural pest on cruciferous crops. Although very common in agricultural fields in north central Florida, this species is now rather rare in south Florida (Lenczewski, 1980). Walker (1991) reported northward movement in the spring in Florida, but it seems this species is also capable of overwintering as far north as Canada (Scott, 1986).
Vanessa virginienis (Drury). Although the American painted lady is a common species in north Florida, this butterfly has been reported by Lenczewski (1980) as infrequent in south Florida (Dade and Broward Cos). The range of this species extends from coast to coast in the U.S., from southern Canada to Colombia, the Canary Islands and Hawaii. It is rare in the Antilles and an occasional vagrant in Europe. This species was not previously known to be migratory, at least to the extent of V. cardui. As found for P. rapae in Gainesville, Florida, V. virginiensis showed northward movement in the spring but individuals were seldom caught in the fall (Walker, 1991). Possibly their movements are not related to temperature since Scott (1986) reports this species as also overwintering in Canada.
Eurema lisa (Boisduval and Le Conte). Scott (1986) reports this species as overwintering in the southeastern U.S. and extending its range north to Canada in summer. At Gainesville, Walker (1991) trapped E. lisa only in the fall when it was moving south. This species is a well-known migrant and has been reported on occasions to fly over the Caribbean, the Atlantic and the Gulf of Mexico in huge swarms (Howe, 1975; Klots, 1951; Lenczewski, 1980).
Eurema daira (Godart). This species overwinters in the southeastern Coastal Plain and spreads north a few hundred kilometers in summer (Scott, 1986). Walker (1991) found no significant movement south in the fall through Gainesville except in the fall of 1985.
Eurema nicippe (Cramer). This species sometimes moves in a seemingly inappropriate direction, northward during fall flights (Walker, 1991) and reportedly overwinters throughout the southeastern U.S. (Howe, 1975). However, it has not been reported in December, January or March in south Florida Dade or Monroe Cos., although it is very common there in the fall (Lenczewski, 1980).
Danaus plexippus CL.). The monarch is undoubtedly the best studied migrant butterfly. Large numbers of monarchs aggregate along the Florida panhandle coast each fall. Most of this eastern population flies west along the Florida coast and down through Texas to overwintering sites in Mexico (Brower, 1977). The migrating monarchs generally fly at altitudes (Gibo, 1986) where flight traps are of little use, although Walker (1991) did trap 15 in ten years and reported a significant southward bias in the fall. Those individuals that enter Florida become physiologically "trapped" when warm
temperatures end their reproductive diapause, and subsequently, migratory tendency (Brower, 1961; T. Van Hook, pers. comm.). There are reports of several small overwintering colonies along the west coast of Florida (Brower, 1961; and pers. com.) but the breeding status of these is unknown. About 20% of female monarchs along the Florida Gulf coast during fall migrations are mated and will oviposit on available milkweeds in Gainesville also during that time (T. Van Hook, pers. comm.). This species is also known to breed in south Florida during the winter months, but the subsequent migratory tendencies of the offspring are not known (Lenczewski, 1980).
Panoquina ocola (Edwards). This small skipper is often found concentrated in large numbers along the Gulf coast with other migrants in the fall (pers. obs.) and has been reported by Walker (1978) as migrating through Gainesville. A mass movement of these skippers in Louisiana was described by Penn (1955). Found through most of the southeast, from Virginia to Florida, west to Texas, Kentucky and Arkansas, it has also been reported as far north as New Jersey (Howe, 1975). There are no records of this species in south Florida (Dade and Monroe Cos.) in January (Lenczewski, 1980).
Lerema accius (Smith). The range of this skipper is from New England to Florida, west to Illinois, Arkansas, Texas and south to northern South America. Howe (1975) reports it as scarce northward but common in the southern states, with records from February to November. Walker (1978) reported some fall southward movement for this species and it is present during every month of the year in south Florida (Lenczewski, 1980).
CHAPTER 2 FLIGHT DIRECTION
Of all insect migratory movements, those of butterflies are the most evident and most accessible to study. Unlike their nocturnal counterparts, the moths, butterflies are highly visible, daytime fliers which move through the boundary layer (Williams, 1930; Baker, 1978). This is the layer of air near the ground where wind velocity is less than the insect's air speed. The thickness of this boundary layer is variable, determined by wind velocity as well as the air speed maintained by the individual (Taylor, 1958; Pedgley, 1982). Most other insect migrants, such as leafhoppers (Taylor and Reling, 1986), noctuid moths, locusts and aphids (Johnson, 1969), travel at much higher altitudes. Radar studies have shown that they are generally flying with the wind (Pedgley, 1982), although they may take off from the ground only when wind direction is favorable (Williams, 1958). Flight within a few meters of the ground, however, affords the individual more control over direction and also the opportunity to cease flying by clinging to vegetation should conditions become unfavorable. This type of flight allows an observer to easily identify species as well as to observe flight direction and behavior under variable environmental conditions.
There are at least eight species of migrating butterflies that pass throughnorth central Florida (Walker, 1978, 1980, 1985ab, 1991) each fall. The four most obvious and abundant species are Phoebis sennae (L.) (Pieridae), the cloudless sulphur; Agraulis vanillae (L.) (Heliconiidae), the gulf fritillary; Urbanus proteus (L.) (Hesperiidae), the long-tailed skipper; and Precis coenia (Hubner) (Nymphalidae), the buckeye. Outside of Walker's studies (1978, 1980, 1985ab, 1991; Walker and Riordan, 1981), mostly based in Gainesville, Florida, there is very little known about the flight directions of these species. The conspicuously yellow cloudless sulphur, especially, has been reported travelling in southeasterly fall flights throughout the southeastern states (Lambremont, 1968; Shannon, 1916; Williams, 1930, 1958; Clark and Clark, 1951; Howe, 1975; Urquhart and Urquhart, 1976; Muller, 1977; Gaddy and Laurie, 1983; Pyle, 1981). Although there are reports of inappropriate northern flights (Gaddy and Laurie, 1983; Muller, 1977), or variable movements, especially along coastlines (Urquhart and Urquhart, 1976), the majority of individuals appear to be navigating toward the Florida peninsula (Walker, 1985a). Walker's trapping and observational data (1978, 1980, 1985ab, 1991) clearly demonstrate that the migrants pass through Gainesville, in north central Florida, each fall and maintain an average flight path approximating 141. However, for most of the rest of the state, flight paths are unknown. My study was undertaken to investigate previously unknown migratory flight directions of the four principal migrant species at various sites throughout the Florida peninsula.
Materials and Methods
During the fall migratory season (September-November) of 1985-1988, flight azimuths for the four migrant species were recorded at a number of locations throughout the Florida peninsula. Sites selected for observations were large, open areas, such as high school football fields or large lawns. Hours for observation were peak flight periods established by Walker (1985a) from 0730 to 1430 HRS LMT. Observation periods were also limited to weather conditions that are favorable to migration and were terminated if more than 50% of the sky was obscured by clouds, temperatures dropped below 21 C or wind speed exceeded 3 m/sec. For each observation, these conditions were noted as described in detail by Walker (1985a): civil time, species, appearance of sun, percentage of blue sky visible, wind speed, wind direction and air temperature. The butterfly passing closest to the observer was selected and a sighting of its flight path was then made from the point where it had passed, to where it disappeared on the horizon, noting the magnetic bearing with a Suunto KB 14 compass. Such observations continued with the next individual spotted and so forth, also as described by Walker (1985a). Magnetic bearings were adjusted for magnetic declination, resulting in true bearings and civil time was converted to local mean time. Walker (1985a) found that mean migratory directions at Gainesville remained approximately 1410 regardless of species, season, time of day or wind. For the purposes of this study, migration data for all dates at a locality were grouped for each species. These data were analyzed using BUTTAZ.BAS, a
computer program designed by Walker (1985a) to calculate mean direction (M.D.), length of mean vector (r) and the 95% confidence interval of M.D. (Batschelet, 1981, Zar, 1984).
A number of individuals of Phoebis sennae were also marked and released during the fall migration season of 1987. This mark-release-recapture took place at two sites in Gainesville, Florida, from September through December. Both sites had gardens with a number of flowering plants that the adults used as nectar sources. The most northern site was Kanapaha Botanical Gardens and the second site was a Gainesville residence located eight miles south of Kanapaha Gardens on Williston Rd. Between my visits, observations of marked individuals were recorded by the resident, Dr. Lincoln Brower, resulting in more complete information at this site. A total of 800 cloudless sulphurs were netted over a period of three days from 10-12 September and given a number on the upper and lower surfaces of the right wing with a permanent felt tip marker, color coded for each site. Sites were visited three times per week until 23 November 1987 and all marked individuals were noted. Observations at the residential site were continued by Dr. Brower throughout December.
Results and Discussion
The mean flight directions during fall migration 1985-1988 through various sites in Florida for Phoebis sennae are shown in Fig. 2-1. The length of the line represents
A=Arcadia n=13,r=.58,161 ,p<.01 AP= Alligator Point n=18,r=.26,68 B=Bartow n=27,r=.88,146,p<.001 BG=Belle Glade n=0 C=Cross City n=51 j=.89,171,p<.001 CB=Crescent Beach n=0 CL=Clermontn=8,r=.45,19I CW=aewiston n=4,r=.77,123 E=Everglades City n=2.r=.51,32 E P= East point n=3,r=.9,46 F=Frostproof n=3,r=.72,92 FC=Fishcreek n=4.r=.53,320 G=Gainesville 141 (Walker, 1985a) GS=Glen SL Mary n=5j=.59,126 H=Hawthome n=61,r=.81.152,pc001 HA=Hastings n=9,r=.33,175 I=Interlachen n=34,r=.94,168,p< 001 TM=Immokalee n=12,r=.62,170 ,p<.01
J=Jasper n=4,r=. 12,162 KB=Keeton Beach n=6,r=.29,330 L=Lake City n=4,r=.80,146
LE=Leesburg n=8,r=.96,133,p<.001 LA=Lake Alfred n=7,r=.74,165,p<.02 LB=LaBeUe n=13,r=.66,153.p<.005 LP= Lake Placid n=4,r=.79.71 N=Newberry n=41x=.80,146,p<.001 OOlga n=13,r=.84,178 JK.001 OK=Okeechobee n=12,r=.63.77.p<.005 P=Palatka n=17,r=.72,143 ,p<.001 PE=Perry n=18,r=.93.144,p<.001 S=Steinhatchee n=46.r-.42,86,p<001 SG=SL George n=2,r=l,54 T=Trenton n=25,r=.72,151,p<.001 TA=Tallahassee n=9j=.32,94 W=White Springs n=2,r=.88,167 YJ=YeehawJunction n=l,r=l,159
Fig. 2-1.--Summary of mean flight directions at various sites throughout Florida for Phoebis sennae during fall migrations, 1985-1988 (direction of arrow=mean flight direction, value given in key for each site; length of line=mean vector r, bold letters=significant r value). ^
the mean vector (r) with sites labeled in bold letters indicating statistically significant vector values for those sites with n >5 using Rayleigh's z test (Zar, 1984). Gainesville data is, as previously reported by Walker (1985a), 141. The observations made in this study along the latitude of Gainesville also show a similar pattern of movement. The inland sites south of Gainesville correspond to the expected flight direction if migrants were continuing along the same track observed at Gainesville. Although not statistically significant, due to small sample sizes at many sites, the movement along coastlines seemed confused and quite variable. Along the Gulf coast, butterflies usually followed the coastline even in inappropriate directions. Previous reports of Phoebis sennae flight directions are southeastward inland and coastwise when near the coast, as reported by a number of different observers over many years (Shannon, 1916; Williams, 1930,1958; Clark and Clark, 1951; Lambremont, 1968; Howe, 1975; Urquhart and Urquhart, 1976; Muller, 1977; Gaddy and Laurie, 1983; Pyle, 1981; Walker, 1985a).
Phoebis sennae is particularly sparse on the east coast considering the southeast bearings of inland migrants. It has been observed by Walker (1985a) that even 90 km from the coast, butterflies in south Georgia fly more southerly than further inland, as though they can detect the coast. This trend is also seen here and no individuals of any migrant species were observed at Crescent Beach, the easternmost site of the transect. The reason for an absence of migrants along the east coast portion of the transect is not clear. Migrating butterflies do tend to follow coastlines possibly using a different orientation mechanism used from inland flight. If north or south coastline flight is not distinguished, then by avoiding the east coast altogether, northward flight during the fall,
possibly resulting in freezing, could be prevented. Dave Baggett has reported (pers.comm.) large numbers of long-tailed skippers flying south along the St. John's River near Jacksonville during some years. This geographical feature may serve as a path when found and, if it can be detected at a distance, may create a funnel drawing migrants away from the east coast areas. Calhoun ej a] (1990) reported some colonization by cloudless sulphurs along rivers that appeared to serve as "dispersal corridors." It has also been observed (see Chapter 3) that this "shadow" or absence of migration along the east coast seemed to widen inland with distance travelled south. Other possible reasons for the absence of these migrating butterflies along the Florida east coast may be the lack of an appropriate habitat or some other adverse environmental conditions found there.
Throughout the state, significant migratory flight directions for Phoebis sennae are usually south or southeast. It is noteworthy that in this species, there appears to be a shift in mean direction immediately north of Lake Okeechobee, as though the migrants "detect" the lake. At the town of Okeechobee, the mean migratory track shifted to the east and most individuals (79%) avoided flying toward the large expanse of water. No cloudless sulphurs were observed along the southern boundary of the lake at Belle Glade, just a short time after seeing many individuals flying through Okeechobee. Again, this seemed to be consistent with observations that the butterflies at the northern border of the lake were going around, rather than flying directly over water. A distinct southward migratory flight was detected as far as Immokalee, just southwest of Lake Okeechobee. This is also the northern limit for much of south Florida's more tropical fauna and flora
and the southern boundary for many northern species. Historically, this is about the point at which damage from a hard freeze is unlikely. It is important to consider the number of visits to a site and the pattern of individual flight directions shown in Fig. 2-2, rather than the mean vector alone. For example, although the mean vector at Steinhatchee (Fig. 2-1) is approximately due east, from Fig. 2-2, it is evident that the individuals had no consensus in migratory direction. The same is true at Alligator Point during the only visit there. The eastward vector in Fig. 2-1 really is a result of south and north flying individuals as shown in Fig. 2-2.
The mean flight direction patterns for the gulf fritillary throughout the state are similar to those described for the cloudless sulphur and are shown in Fig. 2-3. Significantly directed movement for the gulf fritillary continued south to Immokalee, but fewer individuals were observed. Whereas Phoebis sennae is not common along the gulf coast, Agraulis vanillae masses along the panhandle coast in large numbers (pers. obs.). Also, unlike Phoebis sennae. the gulf fritillary showed no evidence of avoiding Lake Okeechobee. Three individuals were observed flying southeast from Okeechobee and others were apparent at the opposite shore in Belle Glade, heading in the same direction. However, this species too avoided Crescent Beach, the most eastern site along the Atlantic Coast.
Whereas some cloudless sulphurs do overwinter in north Florida and buckeyes have also been reported to do so (Walker, 1978; Scott, 1986), this is not usual for the gulf fritillary. The more highly directed flight of the gulf fritillary may be necessary to ensure survival before freezing temperatures occur. The individual flight directions
ARCADIA ALLIGATOR POINT BARTOW
PHP GLADE CROSS CITY
CLERMONT CRESCENT BEACH CLEWISTON EVERGLADES CTTY EASTPOINT
FROSTPROOF F1SHCREEK GLEN ST. MARY HAWTHORNE HASTINGS
JASPER KEETON BEACH LAKE CITY
LEESBURG LAKE ALFRED LABRI .IF, LAKE PLACID NEWBERRY
ST. GEORGE TRENTON TALLAHASSEE WHITE SPRINGS YEEHAW JUNCTION
Fig. 2-2.--Azimuths of Phoebis sennae flight directions taken at various locations in Florida during fall migration 1985-1988. The diameter of the circle represents the number of visits (5 mm = l visit); the length and width of the line represents the number of individuals (2 mm = l, in bold 2 mm =4); the orientation of the line represents the flight direction.
A=Arcadia n=6,r=.98,170 ,p<.001 AP=Alligator Point n=13,r=.17,164 B=Bartown=10,r=.84,174 ,p<.001 BG=BeUe Gladen=0 C=Cross Cityn=21 j=.89,156 .rx.001 CB=Crescent Beach n=o CL=aermont n=8,r=.6,142 CW=Clewiston n=5,r=.92,146 E=Everglades Cityn=6j=.62,109 EP=Eastpoint n=2,r=.99,277 F=Frostproof n=1 ,r= 1,163 FC=Fishcreek n=0 G=Gainesville (Walker, 1985a) GS=Glen St. Mary n=0 H=Hawthome n=27,r=.84,160 ,p<.001 HA=Hastings n=4,r=.84,132 I=Interlachen n=14,r=.96,179 JX-001 IM=Immokalee n=16j=.72,164 ,p<.001
J=Jasper n=4,r=.92 KB=Keeton Beach n=4,r=.7,290
L=Lake Cityn=4,r=.66,154 LE=Leesburg n=5j=.54,163 LA=Lake Alfred n=8,r=.96,155 .rx.001 LB=LaBelle n=2j=.97,164 LP=Lake Placid n=3,r=.99,175 N=Newberryn=16,r=.97,148 jx.OOl 0=01ga n=2j=. 17,253 OK=Okeechobec n=4,r=.96,143 P=Palatka n=12j=.84,172 ,p<001 PE=Perry n=3,r=.94,151 S=Steinhatchee n=13j=.80,137,p<001 SG=St. George n=3,r=.56,141 T=Trenton n=6,r=.34,147 TA=TaJJahassee n=9j=.32,146 W=White Springs n=5,r=.58,174 YJ=Yeehaw Junction n=lj=l,159
Fig. 2.3.--Summary of mean flight directions at various sites throughout Florida of Agraulis vanillae during fall migrations, 1985-1988 (direction of arrow=mean flight direction, value given in key for each site; length of line=mean vector r, bold letters=significant r value).
observed throughout Florida during fall 1985-1988 are shown in Fig. 2-4. Like P. sennae. Agraulis vanillae showed a great deal of variation in flight direction along the Florida panhandle coast at sites such as Alligator Point. However, at Steinhatchee, which is further south along the west coast, there was no evidence of the confused flight that P.. sennae had shown and movement was significantly biased southeast.
The summary of mean flight directions of Urbanus proteus throughout Florida are shown in Fig. 2-5. Its migratory flight patterns are similar to those of the gulf fntillary. Also common along the Gulf coast, this species exhibited a great deal of variation in flight at coastal sites. This is evident in the individual flight directions from various sites in Florida shown in Fig. 2-6. For example, at Keeton Beach, movement was in every direction. Oddly, a similar variation in flight was noted at Yeehaw Junction, an inland location just north of Lake Okeechobee. However, a few miles directly south, at Okeechobee, the flight of individuals was very consistently due south. It appeared they were heading out across the water, but just south of the lake, at Clewiston and Belle Glade, there were no sightings of incoming individuals. It is possible that the long tailed skipper makes shorter, more localized flights than the other two species discussed at inland sites. This may also account for a great deal of variability in flight directions.
Very few individuals of Precis coenia were recorded throughout the state during fall migrations. This is consistent with Walker's (1991) findings from flight traps in Gainesville that most of the movement is northward in the spring. In this study, of the
ARCADIA ALLIGATOR POINT BARTOW BELLE GLADE CROSS CITY
CLERMONT CRESCENT BEACH CLEWISTON EVERGLADES CITY EASTPO
FROSTPROOF FISHCREEK GLEN ST. MARY HAWTHORNE HASTINGS
INTERLACHEN JASPER KEETON BEACH LAKE CITY
LEESBURG LAKE ALFRED LAB ELLE LAKE PLACID NEWBERRY
OLGA OKEECHOBEE PALATKA
ST. GEORGE TRENTON TALLAHASSEE WHITE SPRINGS YEEHAW JUNCTION
Fig. 2-4.Azimuths of Agraulis vanillae flight directions taken at various locations in Florida during fall migration 1985-1988. The circle diameter represents the number of visits (5 mm = l); the length and width of the line represents the number of individuals observed (2 mm = l); the orientation of the line represents the flight direction.
A=Arcadia n=9,r=.81,168 AP=Alligator Point n=0 B=Bartow n=8,r=.66,156 BG=Belle Glade n=0 C=CrossCity n=12,r=.82,147 CB=Crescent Beach n=0 CL=aermont n=3,r=.82,147 CW=Clewiston n=0 E=Everglades City n=6j=.48.121 EP=Eastpoint n=0 F=Frostproof n=2,r=.94,143 FC=Fishcreek n=2,r=.64,318 G=Gainesville (Walker, 1985a) GS=Glen St. Mary n=0 H=Hawthome n=6,r=.89,157 HA=Hastings n=3j=.31,338 l=Interlachen n=3,r=.67,135 IM=Immokalee n=llj=.68,177
J=Jasper n=l,r= 1,152 KB=Keeton Beach n=22,r=.28,168 L=Lake City n=0
LE=Leesburg n=12j=.89,145 LA=Lake Alfred n=2j=l,347 LB=LaBeUe =7,r=.53,215 LP=Lake Placid n=lj= 1,277 N=Newberry n=l,r= 1,280 0=01ga n=5,r-87,168 OK=Okeechobee n=6j=.97,186 P=Palatlca n=o PE=Perry n=2j=l,162 S=Steinhatchee n=23j=.91,120 SG=St George n=0 T=Trenton n=3,r=.98,152 TA=Tallahassee n=0 W=White Springs n=0 YJ=Yeehaw Junction n=13,r=.55,144
Fig. 2-5.-Summary of mean flight directions at various sites throughout Florida of Urbanus proteus during fall migrations of 1985-1988. (The direction of the arrow=mean flight direction, value given in key for each site; the length of line=mean vector r; bold letters=significant r value).
ARCADIA ALLIGATOR POINT BARTOW BELLE GLADE CROSS CITY
CLERMONT CRESCENT BEACH CLEWISTON EVERGLADES CITY EASTPOINT
FROSTPROOF FISHCREEK GLEN ST. MARY HAWTHORNE HASTINGS
Q 0 O
IMMOKA I.I EE INTERLACHEN JASPER KEETON BEACH LAKE CITY
LEESBURG LAKE ALFRED LAB ELLE LAKE PLACID NEWBERRY
OLGA OKEECHOBEE PALATKA PERRY STEINHATCHEE
ST. GEORGE TRENTON TALLAHASSEE WHITE SPRINGS YEEHAW JUNCTION
O G) O O
Fig. 2-6.~Azimuths of Urbanus proteus flight directions during fall migrations, 1985-1988. The circle diameter represents the number of visits (5 mm=l); the length and width of the line represents the number of individuals (2 mm=l); the orientation of the Itine represents the flight direction.
two sites where there were more than two individuals recorded, the movement was southward. The only location where buckeyes were sighted south of Gainesville was in Bartow, central Florida and those few individuals were headed north.
In two instances, cloudless sulphurs were recaptured at some distance from the mark-release sites in Gainesville. These are the first instances of an individual Phoebis sennae being tracked over a relatively long distance and time. The first individual had been marked at Kanapaha Gardens and was seen again feeding at flowers at the Williston Rd. residence, eight miles south of there, ten days later. The second individual had been marked at the Williston Rd. residence and was recaptured in Bronson, Florida, 51 km southwest of Gainesville, 14 days later. This cloudless sulphur had evidently taken refuge in some potted plants during a cold evening and had been brought into the house. In both instances, the butterflies travelled a much shorter distance than would be expected from the elapsed time. It has been calculated (Arbogast, 1966; Balciunas and Knopf, 1977) that migrating cloudless sulphurs or gulf fritillaries go 15-20 km/hr under ideal conditions. Thus, such a journey could easily be completed in one day.
CHAPTER 3 NUMBERS OF MIGRANTS
Each fall, a migratory mass of butterflies comprised of at least eight species from four families, passes through north central Florida. The most conspicuous of these migrants is a pierid, Phoebis sennae (L.), commonly known as the cloudless sulphur. This butterfly has been reported throughout the southeast during fall migratory flights which are usually toward the Florida peninsula (Walker, 1985a). The three other principal migrant species are Agraulis vanillae (L.) (Heliconiidae), the gulf fritillary; Urbanus proteus (L.) (Hesperiidae), the long-tailed skipper; and Precis coenia (Hiibner) (Nymphalidae), the buckeye. Large numbers of these species have been noted along the Florida Gulf coast (Urquhart and Urquhart, 1976) as well as passing through north Florida regularly (Walker 1985ab, 1978, 1980, 1991).
It has been difficult to estimate the size of this migratory flight. Walker (1991) has operated traps for more than ten years in Gainesville, Florida, but the magnitude of migration had not been monitored at any other location. In order to accurately estimate the total numbers of butterflies moving south into central Florida each fall, the density of migratory flights must be monitored at other locations along the latitude of Gainesville (29.65). Walker's (1991) long term trapping results from Gainesville could then be
projected to estimate the absolute numbers moving south each fall. The goal of this study was to construct a "migration profile" of density estimates for the two principle migrants, Phoebis sennae and Agraulis vanillae. along a transect across north central Florida at the latitude of Gainesville. This cross section through the migratory stream will then be used to estimate the total fall migration of Phoebis sennae and Agraulis vanillae into peninsular Florida.
Materials and Methods
Latitudinal Pole Counts
In the fall of 1986, a counting system previously used by Walker (1985a) was used to estimate the numbers of migrants moving through various sites along the Gainesville, Florida, latitude (29.65). Three 3-meter lengths of PVC pipe were used with the second pipe set in the ground 15 meters from the first, and the third, 30 meters beyond the second. All three pipes were along a straight line at 51-231, perpendicular to the average flight track of 141 established in Gainesville by Walker (1985a). All observations were made in large, clear areas (i.e. athletic fields or lawns) to avoid any vegetative or structural influence on the paths taken by migrants. The air temperature (C), percentage of clear sky and sun exposure (b= bright, h=hazy, o=obscured by clouds, p=disk obscured, but position easily determined 5) were noted. Wind speed was measured with a hand held pith ball anemometer and direction was noted. The numbers of individuals of the four principal migrant species were counted crossing between two poles for three five-minute periods with a one minute break between each
observation period. Cloudless sulphurs were easily observed at 45 meters and therefore individuals flying between the first and third poles, were counted. Long-tailed skippers, gulf fritillaries and buckeyes were less visible and were counted if within 15 meters (between the first and second poles). Counts were made at least once every three weeks from 1 September to 2 November 1986 at most of the ten sites (see Table 1.1 for assigned weeks and Fig. 1-1 for location of sites). This period has been shown by Walker (1991) to include >95% of the total migration through Gainesville.
Observations were adjusted for time of day differences by using a "time-of-day" factor, T, from Walker's (1985a) observations of each species' flight variation throughout the day. The average for each half hour was calculated from three days of observation (3 October 1982 and 4, 11 October 1983) by Walker of percent individuals flying throughout the day. This average percent of individuals flying was then multiplied by the T factor which would adjust up to 100% for that time period. A 15 minute adjustment to T was made by averaging two adjacent half hourly observations. For example, if 6.4% of the day's total was the average seen flying at 1000 HRS (EDT) (counts made between 0945-1014) then T would be 15.6. An average of 5.9% flying at 1030 HRS (counts between 1015-1044) results in a T of 16.9. To adjust for 15 minute periods, the average percent flying for both half hourly observations was averaged with a result of 6.15, the resulting T factor for observations at 1015 HRS would then be 16.3. Observations were grouped according to their closest time period in 15 minute intervals and multiplied by the appropriate T factor. The west-east transect sites were approximately 23 kms apart and extended the width of the state from Steinhatchee to
Crescent Beach, a distance of about 200 km. All results of pole counts were summarized by averaging the net numbers of individuals/m/minute from samples made at each site throughout the season.
During the fall migrations of 1986, another technique was devised to estimate the net migration across west/east transects. Counts of north and south flying butterflies were made while driving along selected roads running west to east. This method covered entire transects rather than the IS or 45 m samples made during the pole counts. Transects with low vegetation and therefore good visibility were selected. The major transect was at the latitude of Gainesville, through the same sites as the pole counts. Ideally, the entire transect was travelled roundtrip once per week, with half the counts completed on a particular day. Only cloudless sulphurs and gulf fritillaries were included in the drive counts because they were easily identifiable at a distance. To be counted, a butterfly had to at least have reached the centerline of the road before the car passed. A speed of approximately 97 kph was maintained and counts were recorded for every 16 km of transect from which the net number of indiv/min was calculated. Standard time was noted at the beginning and end of each segment of transect, as was weather. Weather conditions recorded were the percentage of blue sky visible, the appearance of the sun (b=bright, h=hazy, o=obscured, p=disk obscured, but position easily determined 5), air temperature (C), wind direction and speed (m/sec with a hand held anemometer).
During this study, the analysis of the data presented several problems incurred by the distance and time required for sampling and daily and seasonal changes in the density of migration. Walker and Riordan (1981) have shown that by sampling only on days that meet certain weather criteria, day-to-day fluctuations in numbers can be minimized. Driving periods were limited to weather conditions favorable to migration and were terminated if more than 50% of the sky was obscured by clouds, the temperature dropped below 21 C or wind speed exceeded 3 m/sec. There are however, hourly differences in migration throughout the day with peak flight times usually from 1000 to 1400 HRS EDT (Walker, 1985). Since it was not possible to sample simultaneously along the transect sections, it was necessary to accommodate the daily changes in peak flight periods. Adjustments for time of day differences were made to the actual net numbers observed during drive counts (D) by multiplying with the "time of day" factor, T (see explanation for pole counts above).
As a result of time limitations, unpredictable weather changes and other occurrences during sampling periods, not all segments along a transect were always visited on a particular sampling trip. To equalize for the differences in sample number from each transect segment, the time adjusted counts (DnT) were then transformed relative to a time corrected "base" segment (DjT) which was always sampled during each trip. For the western portion of the transect (Gainesville-Steinhatchee), this base was the Gainesville-Newberry segment. For the eastern portion of the transect (Gainesville-Crescent Beach), the base was the Gainesville-Hawthorne segment. The net number of
individuals observed along each transect segment was converted to a proportion of the net number observed along the base segment = DnT/Dj.
Another important influence on the amount of migratory movement observed is a seasonal factor (Walker, 1991). Since the drive counts were made over a period of several months, numbers had to be adjusted for season. In addition, partial profiles from different drive counts had to be combined to estimate the entire trans-state profile. Walker (1991) has collected phenological information on the migration pattern in Gainesville, Florida for more than ten years. His two permanent traps sample continuously throughout the season and have been shown to reflect migration patterns reliably. To eliminate the effects of season and to unify the east and west transects, Walker's trap catches were used to quantify the "strength" of migration for each day that drive counts were made. The net numbers of gulf fritillaries and cloudless sulphurs captured in Walker's two traps were summed for the period during which drive counts were made that year and a mean net number was calculated. The daily net number caught in his traps was then divided by the mean net number to produce a factor that would adjust to the mean net number. The reciprocal of this factor was called the migration index, I, for that day. If the migration was particularly strong, the migration index was correspondingly low. The adjusted data from drive counts was then multiplied by the migration index appropriate to the day that sampling occurred. In this way, the high numbers observed during the peak migratory season were reduced and the more sparse, late or early migration was augmented, in effect eliminating the seasonal
difference. The results of these calculations yielded the estimated migration profiles for 1986, 1987 and 1988.
Dn = NET # INDIVIDUALS/MINUTE FROM DRIVE COUNTS T = TIME OF DAY ADJUSTMENT
D, = NET # INDIVIDUAI^/MINUTE AT BASE SITE (TIME CORRECTED) I = MIGRATION INDEX
ESTIMATED MIGRATION PROFILE = I (DnT/D,).
The summary migration profiles were then derived from the median segment-by-segment values of the yearly estimated profiles for each species. These cross sections of migration density through north central Florida were used to estimate the annual number of P. sennae and A. vanillae butterflies flying into central Florida by applying the migration profiles to Walker's (1991) Gainesville permanent trap data.
Results and Discussion
Latitudinal Pole Counts Cloudless sulphur
Pole counts for the fall 1986 cloudless sulphur migration are summarized in Fig. 3-1. The two coastal sites showed no net movement. At Steinhatchee, this was due to equal numbers of individuals flying north and south, and, at Crescent Beach it was due to a lack of migrants. All net movement southward was statistically significant, except for Palatka, also near the east coast, where not many individuals were seen. Peak
Fig. 3-1.--Summary of pole counts showing Phoebis sennae migration (sum of time adjusted net number of individuals/meter/minute/number of visits) in north central Florida along the latitude of Gainesville (29.65). Solid bars represent statistically significant (chi-square, p < .05) southward movement and open bars were sites with no statistical bias (chi-square, p <.05) in net movement. Counts were made during 1 September to 2 November 1986 at ten sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB=Newberry, GNV = Gainesville, HAW=Hawthorne, ITL=Interlachen, PLK=Palatka, HST=Hastings, CRB=Crescent Beach. The number of visits to each site are shown above the bars.
numbers of individuals were counted through the central sites of Newberry, Gainesville and Hawthorne, averaging about .45 indiv/m/min. The highest count was through Gainesville at .57 indiv/m/min and numbers declined toward both the Gulf and Atlantic coasts.
Pole counts taken of the gulf fritillary migration during fall 1986 are shown in Fig. 3-2. All net movement was southward and the peak migration was concentrated slightly more to the west, passing through Trenton, Newberry and Gainesville. Peak numbers were recorded in Trenton at .58 indiv/m/min followed by Newberry at .48 indiv/m/min. Whereas the cloudless sulphur had little or no migration on either coast, the west coast migration for the fritillary was relatively strong (.3 indiv/m/min). Few individuals were counted at Hastings, which was also the only site lacking statistical significance to the southward net movement. At Crescent Beach however, directly along the coast, movement was recorded at .13 indiv/m/min. The greater presence along the coastlines of this species corresponds to the gulf fritillary buildup along the panhandle coast.
Long tailed skipper
The summary for fall 1986 pole counts of long tailed skipper net movement across the west-east transect at the latitude of Gainesville is shown in Fig. 3-3. The greatest number of individuals move south through the central sites of Gainesville-Hawthorne (.21 indiv/m/min), but migration of Urbanus proteus is also substantial along both the Gulf and Atlantic coastlines (. 14 and 13 indiv/m/min, respectively). Centrally
n i i i i l i-1-1-1-1--tj-1
0 16 32 48 64 80 96 112 128 144 160 176 192
Fig. 3-2.--Summary of pole counts showing Agraulis vanillae migration (sum of time adjusted net number of individuals/meter/minute/number of visits) in north central Florida along the latitude of Gainesville (29.65). Solid bars represent statistically significant (chi-square, p < .05) southward movement and open bars were sites with no statistical bias (chi-square, p <.05) in net movement. Counts were made during 1 September to 2 November 1986 at ten sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB = Newberry, GNV=Gainesville, HAW=Hawthorne, rTL=Interlachen, PLK=Palatka, HST=Hastings, CRB=Crescent Beach. The number of visits to each site are shown above the bars.
STN CRC TRN NWB GNV HAW ITL PLK HST CRB
0 16 32 48 64 80 96 112 1 28 144 160 176 192
Fig. 3-3.--Summary of pole counts showing Urbanus proteus migration (sum of time adjusted net number of individuals/meter/minute/number of visits) in north central Florida along the latitude of Gainesville (29.65). Solid bars represent statistically significant (chi-square, p < .05) southward movement and open bars were sites with no statistical bias (chi-square, p <.05) in net movement. Counts were made during 1 September to 2 November 1986 at ten sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB=Newberry, GNV = Gainesville, HAW=Hawthorne, ITL=Interlachen, PLK=Palatka, HST=Hastings, CRB=Crescent Beach. The number of visits to each site are shown above the bars.
located Gainesville, as well as the two coastal areas were the only locations with statistically significant movement (chi-square, p < .05). When the sampling period was subdivided to early, middle and late migration, a single, statistically southward biased peak shifted eastward from Steinhatchee to Gainesville to Crescent Beach, respectively. The total fall migration resolved into three peaks, which may represent three pathways into central Florida. Large numbers of long tailed skippers were counted near the west coast, but the highest counts were in the central region, passing through Gainesville. A third peak in migration began at the St. John's River in Palatka and extended to the east coast. Dave Baggett (pers. comm.) has reported large numbers of JJ. proteus migrating along the St. John's River during the fall of some years. This pattern may reflect some geographical obstacles, visual attractions, habitat preferences or source areas, such as bean fields.
Migration Profiles Cloudless sulphur
Profile estimates of Phoebis sennae migration in 1986 across the west-east transect at the Gainesville latitude are shown in Fig. 3-4 a,b,c,d,e,f,g,h. The peak net southward movement along the western half of the transect (=STN-GNV portion) occurred through the TRN-NWB or NWB-GNV segments on 14 October, during week 7 of sampling. Throughout the first five weeks, the movement along the western transect remained fairly constant and relatively low, <. 2 indiv/min. During week 6, this rate doubled and in week 7, counts in the western segments (Steinhatchee-Newberry) increased dramatically. Along the peak transect segment (CRC-TRN, Fig. 3-4g, west)
Fig. 3-4 a,b,c,d,e,f,g,h.Profile estimates of Phoebis sennae migration (net number of individuals/minute, adjusted for time of day, sample size and season) across a west-east transect at the latitude of Gainesville, Florida (29.65) from 13 September to 28 October 1986.Counts were taken on nine transect segments defined by ten sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB=Newberry, GNV=Gainesville, HAW=Hawthorne, ITL=Interlachen, PLK=Palatka, HST=Hastings, CRB=Crescent Beach. x=unsampled sites.
ADJ NEI # INDIVMIN S ADJ NET # INDIVMIN
A DJ NET # INDIVMIN
MO Ol to M
ADJ NET t INDIVMIN ijjg MO u
11 m 1 t'
SI g -
ADJ NET INDIVMIN 3
M M u
ADJ NET INDI V/MIN 3 ?5 ? oa <7> *> w u
at this time, the net number of individuals flying south/minute was nearly 40, 8 times greater than seen previously. Unfortunately, that was the only drive count through CRB-TRN during the entire fall. During week 9, the GNV-HAW and ITL-PLK segments along the eastern transect also increased, but only to about a third of what was seen on the Gulf coast. On 16 October, the migration through the two Gulf coast segments (STN-CRC, CRC-TRN) was southward and twenty times higher than that of any other segment, at any other time, throughout the season. It is not known if large numbers of migrants regularly pass through these segments. The one additional sample through STN-CRC (Fig. 3-4c, west) and four samples at TRN-NWB (Fig.3-4 a,c,d,e, west) seem to indicate that at least during the early season it is not common. It is possible that such large scale movements result from synchronous emergence of a locally produced generation. Large fields observed in the Trenton area had extensive stands of Cassia obtusifolia. a primary hostplant of cloudless sulphurs in this area.
Similar profile estimates of the cloudless sulphur migration during fall 1987 are shown in Fig. 3-5 a,b. All net movement was southward and greatest along the western transect through GNV-HAW during week 6. This segment had more than 300 indiv/min passing south, but the net catch for Gainesville that day was only one cloudless sulphur, flying south! This resulted in a high migration index that perhaps inflated what was an average day in the HAW-ITL segment. Along the eastern transect, the numbers were highest through the TRN-NWB segment during week 7, with 4.5 indiv/min. This transect was sampled on 17 October, almost one year to the day when extremely large
0 16 32 48 64 80 96 112 128 144 160 176 192
0 16 32 48 64 80 96 112 128 144 160 176 192
Fig. 3-5 a,b.-Profile estimates of Phoebis sennae migration (net number of individuals/minute, adjusted for time of day, sample size and season) across a west-east transect at the latitude of Gainesville, Florida (29.65) from 26 September to 18 October 1987. Counts were taken on nine transect segments defined by ten sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB=Newberry, GNV = Gainesville, HAW = Hawthorne, ITL = Interlachen, PLK = Palatka, HST=Hastings, CRB=Crescent Beach. x=unsampled sites.
numbers had been noted there in 1986. The peak numbers along this western transect in 1987 were only one tenth of what they were the previous year.
The profile estimates showing Phoebis sennae migration along this transect in fall of 1988 are given in Fig. 3-6 a,b,c,d. Each transect consists of a roundtrip made during the same day and close in time (trip 1 and 2). Each trip consistently showed the peak net movement occurred southward through the central, NWB-INT, or western STN-TRN segment and was lowest in the east.
All profile estimates made of Phoebis sennae net movement during the fall of
1986, 1987 and 1988 are summarized as yearly migration profiles in Fig. 3-7 a,b,c. Most transect segments showed a statistically significant (chi-square test, p < .05) net southward movement throughout the sampling period. The exceptions were PLK-HST, where in 1986 and 1988, net movement was not significantly biased and HST-CRB, in
1987, where no individuals were sighted. During 1986 (Fig. 3-7a), the peak southward movement was through the western segment of CRC-TRN, 37 indiv/min, followed by STN-CRC with 14 indiv/min. As was pointed out in the individual 1986 profile estimates (Fig. 3-4g), this was a result of one sample taken on 16 October. If this trip were not included, the central transect segment from Trenton to Palatka would contain most of the southward movement. Peak movement south for the fall of 1987 (Fig. 3-7b) was through the HAW-INT segment at nearly 200 indiv/min, the highest rate recorded through three years of sampling. The secondary peak was through the adjacent segment of GNV-HAW with about 11 indiv/min. Both of these segments showed an average
WEEK 3 :18 SEP
WEEK 4 :25 SEP
1NVHAW m PLK MST CRapNV HAW (TL PLK HST era
Z 1 Z 2
0 16 32 48 64 80 06 0 16 32 48 64 80 96 KILOMETERS
WEEK 5 : 29 SEP
iTN CRC TIM NWI OMV (TN CMC TUN NWI ONV
WEEK 6 :12 OCT
3 > Q Z
0 16 32 48 64 80 96 0 16 32 48 64 80 96 KILOMETERS
TRIP 1 TRIP 2
III"! 1 V
112 120138 152 188184 201 M12128144160176192
Fig. 3-6 a,b,c,d.-Profile estimates of Phoebis sennae migration (net number of individuals/minute, adjusted for time of day, sample size and season) across a west-east transect at the latitude of Gainesville, Florida (29.65) from 18 September to 12 October 1988. Counts were made on nine transect segments defined by ten sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB=Newberry, GNV=Gainesville, HAW=Hawthorne, ITL=Interlachen, PLK=Palatka, HST=Hastings, CRB=Crescent Beach. x=unsampled sites.
Fig. 3-7 a,b,c.~Migration profiles summarizing estimates of Phoebis sennae migration across a west-east transect in north central Florida along the latitude of Gainesville (29.65). Yearly migration profiles were derived from the median estimate (net number of individuals/minute, adjusted for time of day, sample size and season). Counts were made from 13 September-28 October 1986, 26 September-18 October 1987 and 18 September-12 October 1988 on nine transect segments defined by ten sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB=Newberry, GNV = Gainesville, HAW=Hawthorne, ITL=Interlachen, PLK=Palatka, HST=Hastings, CRB=Crescent Beach. Solid bars represent statistically significant (chi-square, p < .05) southward movement and open bars were sites with no statistical bias (chi-square, p < .05) in net movement. Total number of visits made to each site are shown above the bars.
|STN CRC TRN NWB QNV HAW ITL PIX HST CRB
-210 160 O110 60
0 16 32 48 64 0 S6 112 12S 144 160 17 102
pTN CRC TRN NWB QNV HAW
ITL PIK HST CRB
0 16 32
60 M 112 128 144 160 176 162
0 16 32 46 64 80 86 112 12S 144 160 176 182
migration rate, but traps in Gainesville caught very little. As a result, the high migration index for that day has increased numbers for these segments. If these days were not included, the GNV-HAW segment would still have the peak southward movement at 2.3 indiv/min, followed by the TRN-NWB segment at 1.4 indiv/min. There were no cloudless sulphurs observed during the one trip made through the HST-CRB segment along the Atlantic coast in 1987. The migration profile for fall of 1988 (Fig. 3-7c) showed a peak through HAW-INT of 2.7 indiv/min, nearly twice that of the next highest peak through NWB-GNV (1.4 indiv/min). Movement along either coast was low, with both east and west segments below .5 indiv/min. For all three years, the Atlantic coastal segment of transect, Palatka to Crescent Beach, consistently had the least number of migrants, averaging about .18 indiv/min. Gulf fritillary
The profile estimates of Agraulis vanillae migration during fall 1986 are shown in Fig. 3-8 a,b,c,d,e,f,g,h. On the western transect, movement mostly occurred through the TRN-NWB segment and peaked during week 6 (Fig. 3-8e) with 20 indiv/min. The Gulf coast segment of STN-CRC, was not sampled during this time, but the following week (Fig. 3-8g), it showed peak movement south at 8 indiv/min. Numbers were low along the eastern transect during all the weeks sampled, except for large numbers (23 indiv/min) during week 9 (trip 2, Fig. 3-8h) through INT-PLK. All peak movement through the season occurred within this segment, except for week 9 (trip 1, Fig. 3-8h) where GNV-HAW had nearly 15 indiv/min.
Fig. 3-8 a,b,c,d,e,f,g,h.-Profile estimates of Agraulis vanillae migration (net number of individuals/minute, adjusted for time of day, sample size and season) across a west-east transect at the latitude of Gainesville, Florida (29.65) from 13 September to 28 October 1986. Counts were made on nine transect segments defined by ten sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB=Newberry, GNV = Gainesville, HAW=Hawthorne, ITL=Interlachen, PLK=Palatka, HST=Hastings, CRB=Crescent Beach. x=unsampled sites.
ADJ NET f INDIV/MIN S
ADJ NET INDIV/MIN
S ui t* fJ
ADJ NET f INDIV/MIN
e m m m
_ADJ NET f INDIV/MIN e m m
23 ~ ? 8 *
* i s
M 13 SE
1 w a W
AOJ NET # INDI VMIN
tt S M MO
ADJ NET INDIV/MIN
s 10 S M
W M 0>
ADJ NET f INDIV/MIN m m m
_A0J NET INDIV/MIN
m m M
Profile estimates for the gulf fritillary migration during fall of 1987 are shown in Fig. 3-9 a,b. All net movement was southward and the migratory peak was recorded during week 7 (Fig. 3-9b, west) along the TRN-NWB segment at nearly 15 indiv/min. This increase in numbers was seen one year and a day after the extraordinarily large numbers of Phoebis sennae were noted along the Gulf coast in 1986 (Fig. 3-4g). Numbers of individuals seen along the eastern transect remained low and the only segments with significant movement southward were the central areas of GNV-HAW and HAW-INT.
Profile estimates of gulf fritillary migration during fall of 1988 are shown in Fig. 3-10 a,b,c,d. Each weekly sample is a round trip during the same day and most of the patterns were similar between trips 1 and 2, except during week 3 of sampling (Fig. 3-10a). Although trip 1 of week 3 showed fairly uniform movement across the western transect (all < 2 indiv/min), trip 2 peaked through TRN-NWB with > 10 indiv/min. Other than the TRN-NWB segment on this day, STN-CRC had the next largest number of individuals, during week 3 and week 5 (Fig. 3-10c, trips 1 and 2). The Atlantic coast had relatively low numbers of migrants (< 2 indiv/min) during week 4 (Fig. 3-10b), but in week 6 showed an increase with peaks at HAW-INT (trip 1) and INT-PLK (trip 2).
Yearly migration profiles of Agraulis vanillae movement during the fall of 1986, 1987 and 1988 are shown in Fig. 3-11 a,b,c. The profile for 1986 (Fig. 3-1 la) peaked through the INT-PLK segment (7 indiv/min) with the next peak at TRN-NWB (4.8 indiv/min). During 1987 (Fig. 3-1 lb), the TRN-NWB transect segment also contained the migration peak at 7.5 indiv/min. This rate was nearly three times the number
2 -3 4
5TN CRC TRN
XXX HAW ITL PLK HST CRB
16 32 46 64 80 96 112 128 144 180 176 192
STN CRC TRN NWB GNV HAW ITL PLK HST CRB
0 16 32 48 64 80 96 112 128 144 160 176 192
Fig. 3-9 a,b.--Profile estimates of Agraulis vanillae migration (net number of individuals/minute, adjusted for time of day, sample size and season) across a west-east transect at the latitude of Gainesville, Florida (29.65) from 26 September to 18 October 1987. Counts were made on nine transect segments defined by 10 sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB=Newberry, GNV = Gainesville, HAW=Hawthorne, ITL=Interlachen, PLK=Palatka, HST=Hastings, CRB=Crescent Beach. x=unsampled sites.
2 4 6
WEEK 3:18 SEP
TRIP 1 I I
STN CRC TRN NWB tlNV 5TN CRC TRN IWB ONV
0 16 32 48 64 80 96 0 16 32 48 64 80 96 KILOMETERS
WEEK 5 : 29 SEP
2 4 6 8
TRIP 1 TRIP 2
J 1 1 B STN crc TRN NWBONV STN crc TRN NWB ONV
2 4 6 8 10
WEEK 4 : 25 SEP
TRIP 1 TRIP 2
I INV HAW OL PLK HST CRB wr~mr- 1NV HAW (TL PLK HST CRE
WEEK 6:12 OCT
TRIP 1 TRIP 2
II1 1 INV HAW ITL PLK HST CRB >l||-MV HAW ITL PLK HST CRB
j )M *M 111 im
Fig. 3-10 a,b,c,d. Profile estimates of Agraulis vanillae migration (net number of individuals/minute, adjusted for time of day, sample size and season) across a west-east transect at the latitude of Gainesville, Florida (29.65) from 18 September to 12 October 1988. Counts were made on nine transect segments defined by ten sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB=Newberry, GNV=Gainesville, HAW=Hawthorne, ITL=Interlachen, PLK=Palatka, HST=Hastings, CRB=Crescent Beach. x=unsampled sites.
Fig. 3-11 a,b,c.--Yearly migration profiles of Agraulis vanillae movement across a west-east transect at the latitude of Gainesville (29.65). The yearly migration profiles are from the median estimates (net number of individuals/minute, adjusted for time of day, sample size and season) during 13 September-28 October 1986,26 September-18 October 1987 and 18 September-12 October 1988. Counts were made on nine transect segments defined by ten sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB=Newberry, GNV=Gainesville, HAW=Hawthorne, ITL=Interlachen, PLK=Palatka, HST=Hastings, CRB=Crescent Beach. Solid bars represent statistically significant (chi-square, p < .05)'southward movement and open bars were sites with no statistical bias (chi-square, p < .05) in net movement. Number of visits to each site is above bars.
recorded at the next highest peak, through CRC-TRN (2.7 indiv/min). All southward movement was statistically significant (chi-square, p < .05) except at PLK-HST and HST-CRB during 1987, where numbers were small. During 1988, the migration peak occurred along the Gulf coast through STN-CRC, but at lower numbers than previous years (approx. 4 indiv/min). During all three years, the lowest numbers of individuals moved southward along the east coast.
Numbers of Migrants
A migration profile summarizing the results of 1986, 1987 and 1988 profiles of Phoebis sennae migration is shown in Fig. 3-12. All net movement was southward throughout the fall sampling period (13 Sept-28 Oct). The transect segment with the greatest numbers of cloudless sulphurs moving south was HAW-INT at 2.7 indiv/min, more than double any other segment. The migration rate was fairly even across the other transect segments, averaging about 1 indiv/min, except for PLK-HST (.34 indiv/min) and HST-CRB (.18 indiv/min), where numbers declined approaching the Gulf coast.
The migration profile summarizing Agraulis vanillae movement for 1986, 1987 and 1988 is given in Fig. 3-13. Peak migratory movement for the gulf fritillary was through the TRN-NWB segment at 4.8 indiv/min. The other peak was along the west coast segment, STN-CRC at 3.9 indiv/min. In the case of the gulf fritillary, all three segments with peak movement were along the west coast. Migration east of Newberry was low in general, but the transect segments with the least movement were PLK-HST and HST-CRB, at .81 and.84 indiv/min respectively.
Fig. 3-12.A migration profile summarizing Phoebis sennae migration across a west-east transect at the latitude of Gainesville, Florida (29.65). The profile is derived from a median estimate (net number of individuals/minute, adjusted for time of day, sample size and season) at each site during 13 September-28 October 1986, 26 September-18 October 1987 and 18 September-12 October 1988. Estimates were made along nine segments of a transect defined by ten sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB=Newberry, GNV=Gainesville, HAW=Hawthorne, ITL=Interlachen, PLK=Palatka, HST=Hastings, CRB=Crescent Beach.
QOn I I I I I I i .....
< 0 16 32 48 64 80 96 112 128 144 160 176 192 (KM)
Fig. 3-13.-A migration profile summarizing Agraulis vanillae migration across a west-east transect at the latitude of Gainesville, Florida (29.65). The profile is derived from a median estimate (net number of individuals/minute, adjusted for time of day, sample size and season) at each site during 13 September-28 October 1986, 26 September-18 October 1987 and 18 September-12 October 1988. Estimates were made along nine segments of a transect defined by ten sites: STN=Steinhatchee, CRC=Cross City, TRN=Trenton, NWB=Newberry, GNV = Gainesville, HAW=Hawthorne, rTL=Interlachen, PLK=Palatka, HST=Hastings, CRB=Crescent Beach.
Walker (1991) estimated the size of the seasonally recurring migration through peninsular Florida by extrapolating his flight trap catches from Gainesville using a preliminary estimate of the migration profile from this study. His estimate of the net movement of migrating butterflies each fall was an average of 22.2 million individual cloudless sulphurs and 41.3 million gulf fritillaries. The combined figure of 63.5 million butterflies approaches the estimated number of monarchs reported to overwinter at Mexican sites (Brower, 1985; Calvert el al, 1979). The average rate of net movement for Phoebis sennae and Agraulis vanillae (1986-1988) from the migration profile of each species was used to recalculate Walker's (1991) data (Table 3.1). Walker's (1991)
Table 3.1 Estimates of the Phoebis sennae and Agraulis vanillae fall migration into the Florida peninsula (Lenczewski, 1992 and Walker, 1991). Average absolute net numbers of individuals flying across a north central Florida transect at the latitude of Gainesville (29,65),_
Mean profile height (Lenczewski, 1992) 1.03 2.10
NWB-GNV profile height 1.03 1.43
Profile correction factor 1.00 1.47
Kilometers across state (Lenczewski, 1992) 190.00 190.00
Migrants trapped/6m (Walker, 1991) 800.00 868.00
Est. migrants into cen. Fla. (millions) 25.30 40.40
Est. trapping efficiency (Walker, 1985b) 0.60 0.35
Adj. migrants trapped/6m 1333.00 2480.00
Est. migrants into C. Fla. (millions) 42.20 115.00
calculations were based on an assumption that the average rate of migration across the transect was no less than 60% of the rate at the longitude of his trapping site in Gainesville. In actuality, for Phoebis sennae. his site fell within the NWB-GNV segment of the transect which had a rate of 1.03 indiv/min, identical to the transect average. Walker also used a conservative transect length of 170 km, whereas the actual transect length used in this study is 190 km. Walker's (1991) Gainesville trap yields of 800 individuals/6 m for Phoebis sennae were applied to these data resulting in an estimated influx into central Florida of 25.3 million individuals. With further adjustment for Walker's (1985b) 60% trap efficiency, the size of the cloudless sulphur fall migration into central Florida was estimated as 42.2 million individuals, nearly double the figure given by Walker (1991).
These calculations were repeated using the migration profile information for Agraulis vanillae. also shown in Table 3.1. The average migration along the latitude of Gainesville was 2.10 indiv/min, as compared to the NWB-GNV segment, which had a rate of 1.43 indiv/min, resulting in a profile correction factor of 1.47. After applying Walker's trap catch of 868 individuals/6m and an efficiency of 35%, the final results were 115 million gulf fritillaries, more than double Walker's (1991) estimate of 41.3 million. The total number of these two species of butterflies migrating south through north central Florida in the fall was estimated at 157 million individuals annually. This number does not include the buckeyes or long tailed skippers, which were estimated by Walker as numbering 3.7 and 14.6 million, respectively. Since profiles were not established for these species from drive counts, Walker's estimates cannot be evaluated.
Using his numbers for these other two species, the total migration of these four principle migrants is given as 175 million individuals annually.
Fall 1987 drive counts of Phoebis sennae net migration across west-east transects south of Gainesville, Florida, during fall 1987 are given in Fig. 3-14 a,b,c,d,e,f. The net numbers observed in these drive counts were adjusted only for time of day. Visits to the same transects were approximately 4 weeks apart and week 9 was the only time that the full north-south range of transects was travelled. Statistically significant southward movement is evident south to the Crystal River-Leesburg transect (3) during week 4 (Fig. 3-14a) and continued through week 9 (Fig. 3-14c). It is not known whether individuals are continually moving southward throughout this time or if this is a secondary wave of migrants from subsequent generations. From the pattern observed at Gainesville, there was probably a continual, and gradually diminishing, net southward movement throughout this period. By week 15 (Fig. 3-14f), this northern area no longer showed any significant migration. The only time statistically significant migration was noted south of Crystal River-Leesburg was during week 7 at the Zolfo Springs-Sebring transect. As is seen with results of trap catches (discussed in Chapter 4), there was no evidence of large scale migratory movement through south central Florida. Unless the migratory track deviates from the mean flight direction observed at Gainesville, the migration appears to diminish south of Leesburg. Small numbers of individuals may be moving farther south, perhaps with a second generation augmenting populations around the Zolfo Springs-Sebring area. Drive counts made during fall 1987 depicting the net migration of Agraulis vanillae net migration across various west-east transects south of
Fig. 3-14 a,b,c,d,e.f.--Drive counts of Phoebis sennae migration (net number of individuals/minute, adjusted for time of day) across various west-east transects south of Gainesville, Florida during fall 1987. Transects with a southward or northward net migration of statistical significance (chi-square test, p < .05), are shown with half filled circles weighted in the appropriate direction. Net movement, either north or south, that is not statistically significant (p < .05), is depicted with open circles. Lines with no circles represent transects where there were no individuals seen. Number of individuals and percentage flying south are given in parentheses after each transect.
1. GULF HAMMOCK-KENDRICK
3. CRYSTAL RIVEfl-LEESBURG
4. W1LDWOOD-NEW SMYRNA
6. BROOKSV1LLE-HILL N DALE
7. BAYONET POINT-SAN ANTONK) 8 WESLEY CHAPEL-ZEPHRHILLS
8. LAKELAND-HAINES CITY
10. BARTOW-LAKE WALES
11. FORT MEADE-FROSTPROOF
12. ZOLFO SPRINGS-SEBRING
13. BEE RIDGE-LAKE PLACID
q-Q south bias
- north bias
o-o no bias
GNV LAT (28.65) NONE SEEN
7 MM =20 KM
Gainesville, Florida are shown in Fig. 3-15 a,b,c,d,e,f. A pattern similar to P. sennae is seen for this species which also demonstrated a statistically significant southward movement from Inglis-Belleview through Crystal River-Leesburg during week 4. Subsequently, the farthest significant southward movement seen occurred during week 7 through Bartow-Lake Wales. Again, since the more northern transects were not sampled at this time, it is not known if southward migration continued there during this period, but presumably it did. Significant migration was still seen through the north central transects during week 9 and, from the patterns seen along the Gainesville transect (Fig. 3-11), it is likely that this continued. No individuals were observed on any trips south of Gainesville (Fig. 3-15 d,e,f) after this time and Bee Ridge-Lake Placid also showed no significant migration throughout the season for this species.
The area known as the Green Swamp is located between transects 5 and 7. It is a large area maintained as wilderness, with some grazing areas, for cattle. The wet areas along the Withlahochee River support a variety of flowering plants, especially in the disturbed habitats where cattle are allowed to graze. Large numbers of migrants have been sighted there during some years, particularly in late November (H. Nigg, pers. com.). This largely unexplored area may contain many resources and provide an ideal environment for adult winter survival. There is evidence from Gainesville (Chapter 4), that some cloudless sulphur individuals will remain in an area with nectar sources for several weeks.
Fig. 3-15 a,b,c,d,e.f.-Drive counts of Agraulis vanillae migration (net number of individuals/minute, adjusted for time of day) across various west-east transects south of Gainesville, Florida during fall 1987. Transects with a southward or northward net migration of statistical significance (chi-squared test, p < .05), are shown with half filled circles weighted in the appropriate direction. Net movement, either north or south, that is not statistically significant (p < .05), is depicted with open circles. Lines with no circles represent transects where there were no individuals seen. Number of individuals and percentage flying south are given in parentheses after each transect.
1. GULF HAMMOCX-K END RICK
3. CRYSTAL RIVER-LEESBURG
4. VflLDWOOD-NEW SMYRNA
. BROOKSVILLE-WLL N DALE 7. BAYONET POINT-SAN ANTOMO
I. WESLEY CHAPEL-ZEPHRHJLLS . LAKELAND-HAINES CITY
10. BARTOW-LAKE WALES
II. FORT MEADE-FROSTPROOF
12. ZOLFO SPRINGS-SEBRING
13. BEE RIDGE-LAKE PLACtD
Q-Q SOUTH BIAS
- NORTH BIAS
O-O "O BIAS
GNVLAT (29.85) NONE SEEN
7 MM-20 KM
CHAPTER 4 PHENOLOGY OF MOVEMENT
Many migratory insects, such as aphids and locusts, are greatly dependent on wind and weather conditions to direct their flight once they are airborne (Pedgeley, 1982; Taylor, 1958). These insects, as well as nocturnal moths, fly long distances at high altitudes requiring the use of sophisticated, expensive technology, such as radar, for study (Rainey, 1951; Schaefer, 1976) or extensive mark-release-recapture programs (Li t aj, 1964; Showers s al, 1989). Butterflies are particularly suitable as subjects for the study of insect migration because they are daytime fliers and move within a variable boundary layer, usually within 4-6 meters of the ground (Edwards and Richman, 1977). Wind speeds within this layer are low enough to allow individual control of flight direction (Johnson, 1969). As the butterflies move near ground level, an observer can identify species as well as note behavior and preferred flight tracks. Migrating butterflies typically continue their flight on a linear track and will fly up and over obstacles rather than around (Williams, 1930). This behavioral characteristic can be exploited to selectively capture migrants for research purposes and to quantify movement by using flight traps. Butterflies also have large wing surfaces which make it relatively easy to conspicuously mark individuals for recapture (Walker and Wineriter, 1981).
Despite all these advantages, and, probably because they are not agriculturally significant pests, the often spectacular movements of these insects have been largely ignored. A few workers (Arbogast, 1965, 1966; Balciunas and Knopf, 1977) have made observations over short periods of time and space, but it has been difficult to simultaneously monitor migrations over distance. Through a massive tagging campaign, Urquhart and Urquhart (1976) succeeded in determining the destination of Danaus plexippus (L.), the monarch butterfly, in the Mexican mountain ranges. The migrations of the great southern white, Ascia monuste L., along the Florida east coast were detailed by Nielsen and Nielsen (1952; Nielsen, 1961) over a period of at least 20 years. In Germany, Roer (1959, 1961a, 1961b, 1962, 1968, 1969, 1970) has described the movements of Aglais urticae L., Inachis io., Nymphalis antiopa and other European butterflies, while Baker (1968ab, 1969) has studied Pieris rapae in England. For more than ten years, Walker (1978, 1980, 1985ab, 1991) has maintained directional flight traps in Gainesville, Florida that continuously monitor the annual migrations of eight species of butterflies through this area. This type of data collection eliminates observer bias, time of day or seasonal influences and permits correlation with environmental factors (Walker and Riordan, 1981). A continuous, long term documentation of migration in this manner is unique and has provided reliable information about the species composition, phenology and size of the migration, at least through Gainesville.
There are three species that comprise most of the migration through north central Florida in the fall (Walker, 1991): Phoebis sennae (L.) (Pieridae), the cloudless sulphur; Agraulis vanillae (L.) (Heliconiidae), the gulf fritillary; and Urbanus proteus
(L.) (Hesperiidae), the long-tailed skipper. Another important migrant in Florida is Precis coenia Hubner (Nymphalidae), the buckeye, but most of its migration occurs northward in the spring with the southward fall movement being relatively small. This pattern is the reverse of what is seen for the other species listed. Together, these four species make up more than 85% of Walker's (1991) total trap catch. Walker (1991) captured at least 100 individuals of five other species in his ten years of trapping. Pieris rapae (L.) (Pieridae) and Vanessa virginiensis (Drury) (Nymphalidae) showed significant northward movement in the spring, but were seldom caught in the fall. Eurema lisa (Boisduval & LeConte) (Pieridae) had significant southward movement through Gainesville in the fall, but was never caught in the spring. Another pierid, Eurema daira (Godart) moved south in the fall of one year, but showed no statistical bias in either direction over all years combined. Finally, the pierid, Eurema nicippe (Cramer) was found in some years to move northward in the fall, a seemingly inappropriate direction. The monarch passes primarily along the gulf coast to Texas in its migrations south (Brower gt al, 1985) although, a few individuals do move through north central Florida (Walker, 1991). Since this species flies at a higher altitude than others (Gibo, 1986), they are seldom captured in flight traps.
Although the movements of these species have been well documented in Gainesville by Walker, migratory activities in the rest of the state are virtually unknown. In particular, the timing, or phenology, of the migration south of Gainesville is not known, nor are the characteristics of the migratory front, as far as density and species composition. Arbogast (1966) and Balciunas and Knopf (1977) determined the flight
speed of migrating gulf fritillaries and long tailed skippers to be approximately 12-22 kph under ideal conditions. Weather permitting, peak migration times are usually 0900-1600 HRS LMT (Walker, 1985a) daily, suggesting that individuals may travel as much as 100 km/day. If the same individuals that pass through Gainesville during the peak of migration move directly to lower latitudes, a similar pattern of peaks would be expected at sites to the south. However, if peaks lag to the south by about 3 weeks, the approximate generation time for gulf fritillaries (Arbogast, 1965, 1966), movement south could be in a stepwise, generational manner.
Materials and Methods
Longitudinal Pole Counts
During 1986, pole counts (see description of methods in Chapter 3) were initially made at Gainesville and five sites south, Clermont, Bartow, Arcadia, LaBelle and Immokalee (see Chapter 1, Fig 1-1 for location of all sample sites). The primary purpose of these counts was to get some information on when migration began and how far it extended south of Gainesville before planning more extensive data collection with traps. Counts were made from 13 September to 2 November 1986, at approximately two week intervals (see Chapter 1, Table 1-1 for dates of all sampling). This period was figured by Walker (1991) to include 95% of the total seasonal migration through Gainesville. The six sites along the north-south transect were spaced approximately 62 km apart, a total distance of 372 km. All net numbers were adjusted as explained previously (Chapter 3) for time of day differences (T) using Walker's (1985a)
observations of daily flight periodicity. These adjusted net numbers were averaged to yield a seasonal summary for each site along the north-south transect.
Portable Flight Traps
For the purposes of this study, a portable flight trap was designed (Walker and Lenczewski, 1989) to quantify the phenology of butterfly migration along the Florida peninsula. Walker (1985a) has determined that the migratory track through Gainesville, Florida remained relatively constant at 141 annually. A migratory route was projected north and south (321 and 141) from Gainesville and pairs of these flight traps were erected on it at five sites (an average of 105 km apart) from southern Georgia to south central Florida: Valdosta, Georgia; Gainesville, Leesburg, Lake Alfred and Lake Placid, Florida. Migration was monitored with these traps through the fall of 1987 and 1988. The traps were placed in pairs along an ENE-WSW line (67.5-247.5), one trap facing north and one facing south. These traps, erected approximately perpendicular to the axis of the Florida peninsula and to the mean direction of migration, were left in place from 13 September through 12 December 1987. Butterflies captured by the south-facing trap were flying 321 90; those captured by the north-facing trap were heading 157.5 90. All traps were situated in large, open fields to eliminate the influence of vegetation or other obstacles on the preferred flight tracks of migrants. As the migrant butterflies tried to fly over a main wall of polyester screening, they passed through a narrow slot into a duct that led to two holding cages made of 1/4 inch hardware cloth. Captured butterflies were removed from the holding cages semiweekly throughout the trapping
period. Walker and Lenczewski (1989) estimated that semi-weekly service recovered at least 90% of all individuals entering the traps' collecting cages. After testing other construction materials in the spring of 1988, the traps were modified, primarily to increase durability. The backs of eight polyester traps were replaced with monofilament shrimp netting. Two of the polyester netting traps were left intact. An additional four traps were built of the same design but with all monofilament shrimp netting. All these traps were found to be similar in efficiency (Walker and Lenczewski, 1989), but the monofilament shrimp netting was more durable and easier to work with. In 1988, traps were in place from 28 August through 26 November.
The sites were 157, 100, 70 and 92 km apart, respectively, a total of 419 km from northernmost to southernmost site. The types of traps used each year, at each site, are as follows:
1. Valdosta State College, Valdosta, Georgia (30.85 lat, 83.29 long). 1987 two all polyester traps, 1988 four polyester traps with shrimp net backing.
2. Green Acres Farm, University of Florida, Gainesville, Florida (29.65 lat, 82.44 long). 1987 two all polyester traps. 1988 no traps at this location, data used from Walker (1991).
3. IFAS/AREC, University of Florida, Leesburg, Florida (28.90 lat, 81.54 long). 1987 two all polyester traps, 1988 four polyester traps with shrimp net backing.
4. IFAS/CREC, University of Florida, Lake Alfred, Florida (28.00 lat,
81.44 long). 1987 two all polyester traps, 1988 four all shrimp net traps.
5. Archbold Biological Station, Lake Placid, Florida (27.45 lat, 81.23 long). 1987 and 1988 two all polyester traps. During the fall of 1987, two all polyester traps were in operation at all five sites, one facing north and one facing south. In fall of 1988, four polyester traps with shrimp net backing were used in Valdosta and Leesburg. Four all shrimp net traps were placed at Lake Alfred, and the Lake Placid station had two all polyester traps similar to those used in 1987. During 1988, all stations had two traps facing north and two south, in an alternating pattern, except for Lake Placid, where one trap was positioned north and one south. There were no traps placed at the Gainesville site in 1988 since similar data were available from Walker's (1985b) permanent flight traps. The dates for each week sampled are as given in Chapter 1, Table 1-1.
Results and Discyssipn
Longitudinal pole counts
Pole counts of Phoebis sennae net movement along a north-south transect from Gainesville to Immokalee during 1986 are shown in Fig. 4-1 a,b,c,d,e,f. All statistically significant (p < .05) net movement at the six sites was southward throughout the period sampled. All three weeks sampled at Gainesville (a) showed significant southward movement with the peak occurring during week 2. Of all the sites sampled, Gainesville
.2 0 J A M
1 214I47I4 10
Fig. 4-1 a,b,c,d,e,f.-Pole counts of Phoebis sennae movement (net number of individuals/meter/minute, adjusted for time of day) along a north-south transect through Gainesville, Clermont, Bartow, Arcadia, LaBelle and Immokalee, Florida from 11 September through 7 November 1986. Solid bars represent a statistically significant net movement (chi-square test, p > .05), shaded bars represent net movement which was not significantly biased (chi-square test, p < .05). x=no counts made.
had the greatest numbers moving south (1 indiv/m/min) during peak migration, well over three times that of the next largest movement at Clermont (.25 indiv/m/min). At Clermont (b), which is located approximately 136 kms south of Gainesville, the cloudless sulphur did not achieve a statistically significant southward movement until five weeks later, during week 7. The rest of the sites had no significantly directed movement, except for LaBelle, where cloudless sulphurs also migrated southward (.2 indiv/m/min) during week 7.
The summary of Phoebis sennae pole counts along the transect from Gainesville south is shown in Fig. 4-2. Gainesville had the highest counts with an average of about .6 indiv/m/min passing southward. Clermont and LaBelle, the only other sites with significant southward movement, had a rate of about 1 indiv/m/min. Flight trap catches
Weekly percentages of total net trap catch of Phoebis sennae during the fall of 1987 at five sites along a north-south transect through Gainesville are shown in Fig. 4-3 a,b,c,d,e. All statistically significant net movement was southward for this species, falling between weeks 4 and 12. The early weeks were not sampled completely at any of the sites, and the week in which half of the total net catch was achieved is probably not a reliable estimate of mid-migration point. The Gainesville data (b) were corrected for this by using information taken from Walker's (1991) permanent traps near the same location, shown in Fig. 4-4c. In the case of the cloudless sulphur, the percent total net catches for weeks 5-13 were reduced by 33% (Walker's trap catch during the first 4 weeks). The appropriate adjustment was applied in a similar manner for the other three
Fig. 4-2.Summary of pole counts of Phoebis sennae (average net number of individuals/meter/minute, adjusted for time of day) through Gainesville, Clermont, Arcadia, LaBelle, and Immokalee, Florida during 11 September through 8 November 1986. Numbers below the bars represent total number of counts at that site.
GAINESVILLE PTP#&5 WALKER(1991)
1 > I 4 I I 7 10 tl U It 14 11
WEEKS LAKE ALFRED
1 2 7 10 11 12 It 14 II
1 2 3 4 6 7 10 1112 13 14 16
1 2 3 4 7 10 11 12 13 14 16
Fig. 4-3 a,b,c,d,e,f. Weekly percentage of seasonal total net trap catch of Phoebis sennae at Valdosta, Georgia, Gainesville, Leesburg, Lake Alfred and Lake Placid, Florida in flight traps during 13 September through 12 December 1987. Solid bars represent statistically significant southward biased movement (chi-square test, p < .05), shaded bars show nonsignificant (p >.05) southward movement. Clear bars show nonsignificant (p > .05) northward movement. The total net catch (southbound minus northbound) for the period is shown in the upper right corner. The midpoint of migration is designated by an asterisk, x=trap not operating. Total percent migration in b) was reduced by 33% on the basis of data in c).
species. The estimate for mid-migration was either week 5 or 6 at all sites. The close coordination of mid-migration seems to indicate that the same wave of individuals passed through all five areas. The most southern site, Lake Placid, had no net movement, and even though several individuals were seen and two were trapped, there was no indication of migration. In fact, the numbers of cloudless sulphurs were surprisingly low at all other sites, as compared to Gainesville. The reason for low numbers at Valdosta at least could be that since the first three weeks were not sampled, most of the migrants may have passed southward earlier.
The percentage of total net catch of Phoebis sennae captured at the five sites in 1988 are shown in Fig. 4-4 a,b,c,d,e,f,g,h,i. Gainesville data is taken from Walker (1991). All statistically significant movement was southward biased. At Valdosta (Fig. 4-4a), for the cloudless sulphur, catches in traps 1 and 2 were not well coordinated. Trap 2 went out of commission after week 6 and the percentage of total catch was adjusted in relation to the first six weeks of trap 1 (i.e. decreased by 40%). During 1988, traps were in place by week 1, at which time there was no evidence of migration. The mid-migration point for trap 1 was during week 6, and the only other significant southward movement was seen during week 7. Southward migration occurred during the second week for trap 2, but sample size was small and the trap may already have been damaged. Walker's (1991) data from the two permanent traps located in Gainesville (Fig. 4-4 c,d) both show mid-migration occurred during week 7. The two traps at Leesburg and the average of the two traps at Lake Alfred all suggest a midpoint during week 8 at those sites. The only significant southward movement seen from catches at
Fig. 4-4 a,b,c,d,e,f,g,h,i.--Weekly percentage of seasonal total of net trap catch of Phoebis sennae at Valdosta, Georgia, Leesburg, Lake Alfred and Lake Placid, Florida during 28 August-26 November 1988. Data for Gainesville, Florida are from Walker (1991). Solid bars represent statistically significant southward biased movement (chi-square test, p < .05), shaded bars show nonsignificant (p > .05) southward movement. Clear bars depict nonsignificant (p > .05) northward movement. The total net catch (southbound minus northbound) for the period is shown in the upper right corner. The midpoint of migration is designated by an asterisk, x=trap not operating. Total percent migration in b) was reduced by 40% on the basis of data in a).
VALDOSTA TRAP 1
I I I I I T I I 11 11
GAINESVILLE PTP3 (WALKER. 18911 60
WEEKS LAKE ALFRED TRAP 1
60 40 30
VALDOSTA TRAP 2
X X X X X X X
1 1 I I t I ) t mi 11 II WEEKS
GAINESVILLE PTP6 [WALKER. 1891) 60
12346678S10111213 WEEKS LAKE ALFRED TRAP 2
1 1 I 4 I 7 I I M 11 11 11
80 80 40
1 > I 4 I 7 I I 10 11 1] 11 WEEKS
Lake Alfred trap 2 was during week 11. At these four sites, from Valdosta to Lake Alfred, the mid-point of migration for traps 1 and 2 fell between weeks 6-9 and it is found almost uniformly (except for Lake Alfred, trap 2) at a one week difference between sites. Lake Placid had no statistically significant bias in movement at any time during the trapping period, although five individuals were trapped and others were seen.
The percentage of total net trap catches of cloudless sulphurs are summarized for both the 1987 and 1988 fall migration seasons in Fig. 4-5 a,b,c,d,e. The 1988 trap 1 and 2 data were tested for homogeneity and combined before calculating the mean percent total net catch for both years. The first four weeks of Gainesville 1987 and all of 1988 data used in this summary are from Walker's (1991) permanent trap catches as shown in Figs. 4-3c and 4-4 c,d. The migration midpoint occurred between weeks 5 through 7 at all of the four main sites. Lake Placid had no evidence of migration with a total net of only one individual heading south in two years. The net numbers captured at Valdosta (Fig. 4-5a) were surprisingly low, totaling only 70 individuals in two years, and similar to catches at Lake Alfred (Fig. 4-5d). During 1987 alone, Gainesville traps (Fig. 4-3b) captured nearly three times that number. The total net number at Leesburg for the two years was 275. This was unexpected since it was assumed that most individuals begin travelling south through Valdosta and then move through Gainesville. During 1987, the earlier part of the migration may have been missed, but in 1988 the pattern was similar with comparable numbers at Gainesville-Leesburg and Valdosta-Lake Alfred. Subsequent breeding could augment the migratory stream between Leesburg and Gainesville, accounting for the larger numbers caught at these sites. The sites south of
12346678 8 10 1112131416 WEEKS
2 3 4 6 6 7 8 8 10 1112131416 WEEKS
1 2 3 4 6 8 7 8 8 101112131416
2 3 4 6 8 7 8 8 10 111213 1416
Fig. 4-5 a,b,c,d,e.--Summary of percent total net trap catches of Phoebis sennae at five sites along a north-south transect (Valdosta, Georgia; Gainesville, Leesburg, Lake Alfred and Lake Placid, Florida) during 30 August-12 December 1987 and 28 August-26 November 1988. Solid bars represent net southward movement and clear bars show net northward movement. Total net catch (southbound minus northbound) is shown in the upper right corner and x=unsampled weeks. Data for the first 4 weeks of 1987 and all of 1988 Gainesville are from Walker (1991).