BREEDING BIRDS OF TWO WOODLANDS: COMPARISONS WITH AREAS NORTH OF FLORIDA

SIEVERT A. ROHWER Department of Zoology University of Tampa, Florida 33620 (Present address: Museum of Natural History University of Kansas Lawrence, Kansas 66044) and GLEN E. WOOLFENDEN Department of Zoology University of South Florida Tampa, Florida 33620

Florida has witnessed the forays of ornitholo- western boundary of this bay. The Tampa gists since the times of Bartram and Audubon, Bay region has been the scene of three out- but few investigators have remained in one breaks of human St. Louis encephalitis since locality long enough to gain a detailed knowl- 1959, the last and major epidemic of 1962 edge of its avifauna. Consequently the distri- being the second largest recorded in the bution of the breeding birds of Florida, and United States. Virtually all of the approxi- especially the landbirds, is poorly known, and mately 66 square miles of Pinellas County measurements of their densities are almost south of Sawgrass Lake are occupied by the entirely lacking. This report describes the suburbs of St. Petersburg, an area from which results of censuses of two deciduous woodland numerous cases of encephalitis were reported habitats in Florida. We know of only two in the epidemic year of 1962. Since the Saw- other censuses in broad-leafed forests within grass Lake swamp and wet woods support a the state. Since it has been suggested that the high population of mosquitoes, studies of both breeding landbird fauna of peninsular Florida vector and vertebrate populations were con- is impoverished compared’ with that of con- ducted here in an attempt to link the life his- tinental eastern North America (Robertson tories of arboviruses with their hosts. 1955), in the second part of this paper we compare the Florida censuses with certain PROCEDURES ones made farther north in the eastern United Censusing began on 21 January and terminated on 19 Sentember 1964. Durina this neriod 48 visits during States. which both of us were usually-present were made at no greater than weekly intervals; altogether these to- PART I: THE SAWGRASS LAKE STUDY taled approximately 250 man-hours. Several addi- AREA tional visits were made in 1965. In 1964 densities of breeding species were estimated Sawgrass Lake lies in the southern third of by systematically traversing our two plots (though Pinellas County, 4.8 miles north and 1.9 miles they could not be thoroughly worked on any one west of the city hall of St. Petersburg. It is visit) and recording the number and location of in- the only extensive fresh-water swamp and wet dividuals on Xerox copies of aerial photographs. Im- pressions of density gained in this manner for the 96- woods near the city. The area is bounded to acre tract were checked by intensive study of two the north and northeast by pine flatwoods and quadrats wherein all individual birds were counted pasture, and on the other sides by encroaching on every visit. suburbs. Our two ’ study plots of 96 and 30 VEGETATION acres lay in the main body of woods northeast Two major habitats occur in the Sawgrass of the lake; they were bounded to the west by Lake study area, including a red maple (Acer the interior edge of the lake ecotone and on rubrum trilobum) wet woods of 96 acres and the other three sides by canals. a mixed oak ridge of 30 acres. The maple Study of the breeding avifauna of the Saw- woods can be subdivided by its vegetation grass Lake deciduous woodlands was con- into three habitats: 6 acres of large manles. 71 ducted as a part of a larger investigation of acres of small maples, and 19 a&es 0; mixed the ecologies of arbovirus in the Tampa Bay small maples and coastal plain willows (S&x region of Florida. Pinellas County forms the caroliniunu ) . One of our quadrats, 5 acres in

The Condor, 71:38-48, 1969 [381 BREEDING BIRDS OF TWO FLORIDA WOODLANDS 39

TABLE 1. Trees and shrubs recorded in the three subdivisions of the 96-acre maple wet woods.

AVerage Number of Size range JXcXEe~ Modal size trees in inches in inches

Large Maple Area ( 359 transect yards) 24 Red maple 2 -16.5 9.3 10 ( 4 individuals) ( Acer rubrum trilobum) 2 Dahoon 4-6 - (llex cassine) Small Maple Area ( 834 transect yards) 192 Red maple OS-14 3.6 2 ( 35 individuals ) (Acer rubrum trilobum) 9 Swampbay 1 -10 4.5 4 ( 3 individuals) (Persea palustris) 4 Dahoon 1 - 3.5 - - ( llex cassine ) Mixed Small Maple and Willow Area ( 398 transect yards ) 134 Red maple 0.5-12 2.2 2 (28 individuals) ( Acer rubrum trilobum) 1 ( 29 individuals ) 85 Coastal plain willow 1.5- 8.5 3.5 3 ( 35 individuals ) (Salk carolinana) 16 Buttonwood 0.5- 3 1.1 1 ( 11 individuals) (Cephahthus occidentalis) area, was within the 71 acres of small maples, In addition to elderberry the low-growing the other of 3 acres was within the 6 acres of vegetation of the interior of the maple woods large maples. The 30-acre ridge was domi- consisted of Virginia willow ( Zteu virginicu ), nated by oaks, but contained a small, centrally pokeberry (Phytoluccu rigidu), wild coffee located, open sari,,, crest. The areas of these ( Psycotria neruosa), and babypepper ( Riuinu subdivisions were calculated from aerial photo- hum&s). These plants were common but graphs. their numbers were not recorded since they The vegetation of each of the four habitats were either much smaller or less abundant was assessed by yard-wide parallel transects than the elderberry. totaling 1989 yards. We recorded the diam- In the area of mixed willows and small eter at breast height (DBH) of every tree or maples the willows are being replaced by the shrub within the strip. In the frequent cases maples. Forty-one dead willow trunks (not of multiple-trunked willows or maples, each included in table 1) were recorded on the trunk was individually recorded. Other than transect of maple-willow habitat, and all of counting the clumps of elderberry ( Sambucus the living willows were scraggly, having leaves simpsonii) we made no attempt to record only at their tops. Willows were abundant quantitatively the herbaceous vegetation. near the lake (outside of our study area) and The maples in the Sawgrass Lake area apparently are one of the first woody plants grow in low-lying muck soils that were for- to succeed the aquatic herbs; however, as the merly flooded during the summer rainy sea- rotting vegetation turns to muck, maples es- son. Canals, which were dug as a part of a tablish themselves and appear to shade out the mosquito control program, presently prevent willows. flooding of much of the maple habitat. The The maples, though apparently healthy, are trees and shrubs recorded on the 1591 yards probably not reproducing themselves. We of transect passing through these maple-dom- found no saplings in the large maple quadrat inated portions of our plot are summarized in and no seedlings in the entire study area. No table 1. The dominant undergrowth of these bald-cypress ( Tuxodium disticum) or stumps muck-lands was elderberry, 264 clumps of thereof were found in the woodland. Saw- which were recorded. Individuals 10 to 12 grass ( Muriscus jumuicensis) was present only feet high were abundant along the well-illumi- in the wetter areas. It appears to have been nated canal banks where the maples had been more widespread in the past, but is being cleared to permit excavation; those of the eliminated by the drainage program and the shaded interior were much smaller and usually shade of the maples. bore no fruit. Primrose-willow (lussiaea peru- In order of decreasing abundance the ground uiuna) and Caesar weed ( Urena lobutu) were ferns of the muck soils were: royal fern (Os- also common on the canal banks but were not munda regalis), cinnamon fern (Osmundu found in the shaded interior of the loiwlands. cinnumoneu) , and giant fern (Acrosticurn 40 SIEVERT A. ROHWER AND GLEN E. WOOLFENDEN

TABLE 2. Vegetation recorded in the 398 yards of transect crossing the mixed oak ridge.

Number of Size range AV;ynyhzS “,“,“,;tse Plant individuals in inches

Saw-palmetto 31 - - - (Serenoa repens) Grape vines 20 l- 4 1.9 1 ( 12 individuals) (Vic.s sp.) Sweetbay 16 l- 8 3.1 3 ( 4 individuals) (Magnolia virginiana) Live oak 13 2-25 8.0 3 ( 5 individuals) ( Quercus virginiana) Laurel oak 4 ( DBHs 14, 3, 2, and 2 inches) ( Quercus hemisphaerica) Possum-haw 3 (all about 1 inch DBH ) (llex decidua) Black gum 1 (DBH 16 inches) ( Nyssa biflora) ( sp.?) dunaefolium); all formed large, elevated ra- (Torostoma rufum) confined most of their tine tussocks. activities to the ridge, as did Screech Owls The mixed oak ridge was surrounded by the (Otus ado), although in the latter case we maple wet woods. The vegetation of its sandy have fewer supporting observations. The one soils was different and much more complex pair of Red-shouldered Hawks (Buteo Zineatus) than that of the muck-lands. Table 2 sum- and their young foraged throughout the woods marizes its vegetation as recorded in 398 yards surrounding the lake. They were never seen of transect. This transect did not cross the in the nearby open country frequented by white sand crest which was dominated by several Red-tailed Hawks (Buteo jamaicensis). sand-live oak (Quercus geminuta). Other The one pair of Green Herons (Butorides plants present on the mixed oak ridge but not virescens) regularly seen in the study area in our transect included common persimmon spent considerable time elsewhere; the pair ( Diospyros virginianu), cabbage palmetto of Mourning Doves (Zenaidura mucroura) (Sabal palmetto), pignut hickory (Carea also foraged outside of the area but fledged glabra ), cherrylaurel ( Prunes curoZiniunu), at least one young from a nest on the ridge. red mulberry (Morus ruber), tree snarkleberry The few pairs of White-eyed Vireos (Vireo ( Vaccinium arboreum ) , citrus ( Citrus sp. ), griseus) and Yellowthroats (Geothlypk trichas) sumac (Rhus copallium) greenbriar (Smilux we recorded confined most of their activities sp. ), blackberry (Rubus sp. ), and beauty- to the canal borders. berry (Callicarpa americana), Some Caesar Pairs of three species had territories entirely weed also grew around the edge of the ridge within the maple wet woods. In the three- but it was more abundant along the canals. acre large maple quadrat their densities were: One area of the ridge, slightly more than an Carolina Wren, two pairs; Cardinal, less than acre in extent, consisted of a tangle of buck- two pairs; Yellow-billed Cuckoo, one pair. thorn ( Sageretia minutiflora). The predomi- In the five-acre small maple quadrat the nant fern was sword fern (Nephrolepsis densities were: Carolina Wren, two and one cordifolia) which did not form elevated tus- half pairs; Cardinal, just under two pairs; socks characteristic of the ferns growing in Yellow-billed Cuckoo, one pair. Thus the the muck areas. small maple habitat had lower densities per Plant identifications and names follow Small unit area. ( 1933, 1938) and Kurz and Godfrey ( 1962). Cardinals were somewhat more common on RESULTS the oak ridge than in the maple-dominated Only 12 species of birds bred in the entire wet woods; the densities of cuckoos and wrens Sawgrass Lake area. Estimates of their were not perceptibly different in the two densities in the two study plots are listed major habitats. For all 126 acres of the Saw- in table 3. Carolina Wrens (Thryothorus grass Lake study area cuckoos, wrens, and ludoviciunus) , Cardinals ( Richmondena car- Cardinals comprised 93 per cent of the breed- din&is), and Yellow-billed Cuckoos ( Coc- ing birds. Active nests of the Mourning Dove, cynus americanus) nested in the maple wet Carolina Wren, and Cardinal were found. woods as well as on the oak ridge. Blue Jays Wrens built nests in fern tussocks in the small (Cyanocitta cristata) and Brown Thrashers maple area. Although no active nest of the BREEDING BIRDS OF TWO FLORIDA WOODLANDS 41

TABLE 3. Breeding birds of the two broad-leafed woodlands in the Sawgrass Lake study area in 1964.”

30;~kr-ac~g~d 96;+$mm?&;le

a&la1 pairs per actual Species density 100 acres density E”,%

Green Heron (Butorides uirescens) 0 - - Red-shouldered Hawk (Buteo line&us) - Monrning Dove (Zenoiduru macroura) T - 0 - Yellow-billed Cuckoo (Coccyzu.s americanus) 6 20 19 20 Screech Owl (0tu.s do) 2 7 Great Horned Owl (Bubo virginianus) + - + ’ - Blue Jay (Cyanociffu cristata) 5 17 Carolina Wren ( Thryothorus ludovicianus) 15 50 4: m Brown Thrasher ( Toxostoma rufum ) 3 10 V - White-eyed Vireo (Vireo griseus) 0 - canal ecotone only Yellowtbroat (Geothlypis trichas) 0 canal ecotone only Cardinal (Richmondena cardinalis) 15 so 38 40 total density 46 154 106 111 number of species 9 6

*v = visitor; + = ranged outside of study area.

Yellow-billed Cuckoo was observed, a female tyrannus), Boat-tailed Grackles ( Casti& collected on 27 April 1964 contained a fully mexicanus), and Common Grackles (Quiscalus formed egg, and we found several old nests quiscula ), they were unaccompanied. Juve- and shells from hatched cuckoo eggs. Other nile Mockingbirds (Mimus polyglottos) made old nests, apparently those of Brown Thrashers the most dramatic invasion. Our first record were located in the shrubs on the mixed oak was for 17 June, when we counted 8; on 29 ridge. June we recorded 45 alone one canal border- A number of local species regularly visited ing the plot. Over 25 Mockingbirds were the study area. They may be treated best identified on every trip thereafter until early under two categories, including species for August, and additional individuals were over- which adults were visitors, and species for looked, undoubtedly, since we spent little time which juveniles were visitors. Few adult birds along the canals where they conareeated to visited the study area during the breeding feed on the abundant ripening elderberries. season, and only four species were recorded Between 7 July and 5 August 187 fledgling on more than 5 of the 17 trips made between Mockingbirds (plus only two adults) were 15 May and 15 June. These were the Purple netted and color-marked to gain information Martin (Procne subis), five to ten of which on their dispersal from the Sawgrass Lake regularly fed low over the trees around the area. Over 35 sightings of marked individuals edge of the ridge; Great Crested Flycatcher were later reported from nearbv residential ( Myiurchus crinitus), several of which hawked areas (Kale and Jennings 1966). Manv young insects on the ridge and earlv in the season Blue Jays also invaded the Sawgrass area. Six appeared to be seeking nest sites thereon; unattended juveniles were observed on 15 Red-winged Blackbird ( Agelaius phoeniceus), June and, because they were independent. a dozen or more females of which obtained both of us felt that they had not been pro- food in the vegetation along the canals: and duced in the study plot. By 14 July the num- Red-bellied Woodpecker (Centurus carolinus), ber had increased to 50, and over 15 were two or three of which foraged in all habitats recorded on each trip thereafter until the end of the study area. Pairs of the latter three of August. species nested nearby along the edge of the Most of Pinellas County south of the Saw- deciduous woods. Adults of several other grass Lake area is either residential or com- species (including herons, charadriiforms, mercial, and the remaining natural habitats icterids, and the Fish Crow, Corvus ossiiragus) are rapidly disappearing. In the residential were frequently recorded overhead but never, areas total breeding-bird densities are as high or rarely, fed in the plot. as 566-600 pairs per 100 acres, with the In the late summer a remarkable number densities of Blue Jays and Mockingbirds being of young birds invaded the Sawgrass Lake exceeded only by House Sparrows (Passer area from other habitats. None was seen to domesticus) and Mourning Doves. For a few be cared for by an adult, and, except for species these residential areas offer great ex- family groups of Eastern Kingbirds (Tyrannus panses of ideal habitat, and the suburbs are 42 SIEVERT A. ROHWER AND GLEN E. WOOLFENDEN

virtually blanketed by their adjacent terri- More detailed quantitative analyses of the tories. In two years of censusing in study plots breeding birds of any portion of the eastern in these suburbs, we saw no independent deciduous forests are difficult to make. The fledgling Blue Jay or Mockingbird, and we region is geologically old, and clearly delimit- frequently saw young birds being driven away ing physiographic features are absent. Thus by adults (Woolfenden and Rohwer, unpub- the different types of forests grade gradually lished data). Because breeding densities are into one another. Correspondingly, the breed- high and suitable refugia for young birds are ing densities of many individual bird species scarce, it seems that fledglings upon becoming vary between forest types, with most not independent are literally driven out of the simply being present or absent, but instead city as subsequent nesting is begun. The being more or less abundant as one changes Sawgrass Lake wet woods with its low breed- forest types. Bond (1957) has ably demon- ing bird density and abundant food sources strated this phenomenon in southern Wiscon- serves as a refuge for these juveniles. sin by using 64 forest stands for which vege- In late June and throughout July a few tational descriptions had been carefully juveniles were recorded for the following 10 quantified. These stands formed a gradient species (the largest number of individuals from xeric oak forests to mesic sugar maple recorded on any one visit follows the name): (Acer sac&u&urn) climaxes, and the densities Yellow-crowned Night Heron (Nyctanussa of most bird species were found to vary con- violacea) , 1; Red-bellied Woodpecker, 1; tinually along this gradient of forest types. Downy Woodpecker (Dendrocopos pubescent), Furthermore, the density of many species fol- 1; Eastern Kingbird, 1; Loggerhead Shrike lowed normal curves or what appeared to be (Lank ludovicianus), 1; Starling (Sturnus parts of normal curves. For species where an vulgaris), 5; Red-winged Blackbird, 60; Boat- entire curve was present, the range of Bonds’ tailed Grackle, 2; Common Grackle, 3; and forest types apparently contained both their Rufous-sided Towhee ( Pipilo eythrophthal- xeric and mesic tolerance limits (at least for mus), 5. Young Red-winged Blackbirds were southern Wisconsin). The American Redstart recorded only once, on 6 August, when a (Setophuga r&i&a) for example, was nearly flock of over 60 individuals, plus a few adult absent in the most xeric stands but as the females, was seen feeding along one of the forests became increasingly mesic they in- canals. Concentrations of fledgling Brown creased in abundance to a maximum in mixed Thrashers (6+) and Cardinals ( lO+ ) were mesophytic stands. Then, as the forests be- observed on the ridge, but in neither case did came even more moist, their densities began we feel that they came from areas outside the to decline until, in the most mesic climax plot. forests, they were again almost absent. In discussing Bonds’ paper, Mengel (1965: PART II: COMPARISONS WITH AREAS 97) points out that “many species tend to NORTH OF FLORIDA slide‘ ’ up and down the forest continuum” as Robertson ( 1955) was the first to discuss the one changes geographic localities. Thus, a impoverished nature of the breeding landbird species ’ point of maximum abundance on the fauna of Florida. He found support for his continuum of forest types may be expected to contention in a general comparison of several shift as one moves from one region to another. regional studies. In seven areas in the north- It becomes evident that a detailed analysis eastern United States and Ontario an average of bird distribution in the eastern deciduous of 73 passerine species was recorded breeding, forests must take densities of individual species whereas only 56 species were listed for the into account. Unfortunately the vegetational entire state of Florida ( Howell 1932). Robert- descriptions accompanying most censuses are son also noted that the number of breeding not adequate for a fauna1 analysis which passerine species decreases as one proceeds could include densities. south on the Florida peninsula, passing in its In the southeastern United States two major central region the southern terminus of the physiographic provinces, the Piedmont and the breeding range of a number of widespread Coastal Plain, occur south and east of the northeastern species. In extreme southern Blue Ridge Mountains (Fenneman 1936). Florida (Dade and Monroe Counties) Robert- The latter includes all of peninsular Florida. son found breeding records for only 29 pas- Since Norris (1958) found some differences serine species, 9 of which are either no longer in the avifauna of the Piedmont and the present, extremely rare, or confined to isolated Coastal Plain in his study of the distribution segments of suitable habitat. of birds in Georgia, our analysis of the breed-

44 SIEVERT A. ROHWER AND GLEN E. WOOLFENDEN

TABLE 5. Matched-pairs comparisons of species composition (above diagonal) and fauna1 size (below di- aeonal) for nine censuses from the Coastal Plain of mainland United States I-IX) and four from nenin- %, I C~ , sular Florida (X-XII), made from data given in table 4. See text for explanation.

X XIA XIB XII =l

I 89 80 100 67 II 67 60 67 III 100 80 1:: 67 IV V 78 80 67 67 VI 100 100 100 67 VII 100 80 100 67 VIII 78 60 100 67 IX 67 60 100 33 X 60 75 53 69 56 50 !: 69 64 XIA 33 42 29 38 31 28 38 36 XIB 20 25 18 23 19 17 20 23 21 XII 20 25 18 23 19 17 20 23 21 ing birds of peninsular Florida will make species found in the larger; however, the two faunas comparisons with censuses taken only on or may be quite different because of a reduced number of species in the smaller fauna. near the Coastal Plain in Alabama, Georgia, Our second procedure is designed to quantify dif- and South Carolina. ferences in the size of two faunas. This is accom- plished by dividing the number of species in the PROCEDURES smaller of two faunas by the number in the larger. Breeding-bird censuses were available for only five Again the quotient is converted to a percentage, and deciduous forests on the Coastal Plain of these neigh- results may range from 0 to 100, with similarity in boring states. In order to increase our comparative fauna1 size being indicated by higher percentages and sample we included censuses from four areas that dissimilarity by lower percentages. lie just above the Fall Line in the Piedmont, two RESULTS each from Birmingham, Alabama, and Athens, Geor- gia. All of these censuses were composed entirely Inspection of table 4 discloses significant dif- of species typically found on the Coastal Plain, so ferences between the censuses made on the our comparisons are not biased by introducing ele- Coastal Plain north of Florida and those taken ments of the Appalachian avifauna. Table 4 summarizes the data obtained from the on the peninsula. The average number of nine censuses of areas lying north of Florida and the species recorded on the censuses in Alabama, four censuses from peninsular Florida. Following Georgia, and South Carolina was 23; for Norriss’ procedure ( 1951) we have included in table peninsular Florida the maximum was only 16, 4 only those species that constitute five per cent or and the average was a mere 9. Pairs per 100 more of any one fauna. Once a species qualified for the list its percentage was given for all censuses on acres averaged 320 for the localities north of which it accounted for one per cent or more of the Florida (or 268 if the extraordinarily high fauna. In addition to presenting percentage com- hammock census from Damascus, Georgia, is positions, table 4 gives the number of pairs per 100 deleted); for Florida the maximum was 183, acres for each species from which the percentages were derived, and the total density and total number and the average was only 117. In numbers of of species recorded for each locality. (In cases species and total density the nine areas north where plots were censused more than once we used of Florida are similar to other eastern hard- averages, and all figures were rounded to whole wood forests. Kendeigh (1944) reviewed a numbers. ) The census data presented in table 4 have been large number of breeding-bird censuses, pri- analyzed in detail by two comparisons designed to marily from the north-central portion of east- quantify the degree of similarity between faunas. ern United States, and concluded that eastern (Speciesrecorded in very low densities, less than one forests are typically inhabited by 20 to 30 per cent in table 4, were excluded from these com- species, and have densities of 200 to 300 pairs parisons.) The first comparison expresses the degree of similarity in the species composition of two per 100 acres. In both respects Florida is faunas. The method has been discussed in detail by depauperate. Further comparisons of total Simpson ( 1947), and applied to bird distribution densities are not made because so few censuses by Miller ( 1951) and Mengel ( 1965). In this com- with detailed descriptions of vegetation are parison the number of species common to two faunas is divided by the total number in the smaller fauna. available for the southeastern Coastal Plain. The quotient is then converted to a percentage, so The results of our analyses of the data that results may range from 0 to 100. Similarity of shown in table 4, using the two methods ex- fauna1 composition is indicated by higher percentages plained above, are given in table 5. In this and dissimilarity by lower percentages. The compari- son of species composition gives high values when table the measurements of species composition the smaller of two faunas is composed primarily of lie above the diagonal and the measurements BREEDING BIRDS OF TWO FLORIDA WOODLANDS 45 of fauna1 size lie below the diagonal. The peninsular coastal Sawgrass Lake study plots intra- and inter-regional comparisons of Flor- are more similar to the south Florida ham- ida and the three states to the north are mock, whereas the mid-peninsular oak-palm separated by a horizontal and a vertical line. hammock, which lies farther inland, is more We verified our conclusions with the Mann- similar to the census areas to the north. The Whitney U test, a nonparametric test for deter- dissimilarity of these two mid-peninsular study mining whether or not two samples represent areas is explained by the breeding distribu- populations of different central tendency. tions of many continental forms in Florida, Intra-regional differences for Florida (lower which typically breed farther south in interior right rectangle) will not be discussed because Florida than along the coast. Thus the oak- the number of censuses available is too small. palm hammock which was located 27 miles The upper left rectangle contains the intra- inland from Sawgrass Lake had breeding regional measurements of species composition Tufted Titmice (Paws bicobr), Blue-gray (above the diagonal) and fauna1 size (below Gnatcatchers ( Polioptilu cumuka) , Red-eyed the diagonal) for the three states north of Vireos (Vireo olivaceus), and Parula Warblers Florida. These measures establish bases with ( ParuZu ameticanu), all of which breed south which the Florida censuses are compared. of both areas, but only in the interior of the Variation in these intra-regional measures pri- state. marily comes from comparing different types We intentionally excluded from the fore- of broad-leafed forests, which was necessary going analyses certain censuses of the southern because of the scarcity of census data. Coastal Plain north of Florida because either Indexes of species composition within the the habitats were not dominated by broad- northern states ranged from 62 to 100 (with leafed trees (e.g. Audubon Field Notes 20: one exception of 50), with a mean of 80. The 629, census no. 33) or the study plots were censuses taken in are similar too small (e.g., op. cit., 20: 609, census no. (columns X, XIA, XIB ), with a range from 5). One additional study is available from 60 to 100 and a mean of 83, indicating the two Florida, that of a 25-acre longleaf pine (Pinus regions are not significantly different from palustris) -turkey oak ( Que~cus luevti) habitat each other in species composition (P = 0.4472). near Gainesville, which we excluded because The one census from southern Florida with the plot was dominated by pines. In five eight inter-regional measures of 67 and one of consecutive years of censusing, this plot had 33 (column XII) is significantly different from only 6 to 10 breeding species and 54 to 96 the more northern states (P < 0.00946) be- pairs per hundred acres. Six species accounted cause it had a total of only three species, one for five per cent or more of the total breed- of which was the Black-whiskered Vireo ing density during the five years, two of (Vireo altiloquus), a colonizer from the West which are not listed in table 4, the Yellow- Indies. throated Vireo (Vireo fluvifrons) and Moum- Intra-regional comparisons of fauna1 size ing Dove (op. cit., 12: 450, census no. 15; for the states north of Florida range from 67 13: 466, census no. 14; 14: 497-498, census to 100, with a mean of 85. When compared no. 18; 15: 524, census no. 43; 16: 527-528, with these northern censuses the oak-palm census no. 19). hammock (row X) shows a significantly re- duced fauna with a range from 50 to 75 and a DISCUSSION mean of 62 (P < 0.00006). The Sawgrass The preceding data and analysis exemplify Lake censuses (rows XIA, XIB ) and the south the depauperate nature of the breeding land- Florida tropical hammock (row XII) have bird fauna of peninsular Florida. These facts even smaller faunas with their comparisons lead naturally to the question of why an im- ranging from 17 to 42 with a mean of 25 (P < poverished fauna occurs on a subtropical 0.00006). In these comparisons with northern peninsula of a temperate land mass. Robert- areas the two Sawgrass Lake censuses are son (1955) considered the question and pre- significantly different from each other be- sented a twofold explanation. First, he sug- cause Blue Jays and Brown Thrashers nested gested that many of the species breeding in on the mixed oak ridge but not in the adjacent eastern North America are not adapted to the maple woods in 1964, a fact made less sig- climate of peninsular Florida. Second, he nificant by Rohwers’ discovery in 1967 of an suspected that natural colonization of the active jay nest in the small maple portion of peninsula by tropical species, which should be the maple wet woods. adapted to the climate, has been deterred by In these comparisons of fauna1 size, the mid- physiographic barriers. In support of his first 46 SIEVERT A. ROHWER AND GLEN E. WOOLFENDEN

supposition he pointed out that the southern to 90 pairs per hundred acres in urban Illinois breeding limits of many northeastern species (Graber and Graber 1963) perhaps reflect lie in north-central Florida, although present- their recent arrival in peninsular Florida. day physiographic features that might prevent Thus, despite a lack of significant geographic them from ranging farther south appear to be barriers and a general north-south migratory absent. Furthermore, he noted that at least pattern, features which might have encour- four birds, the Sparrow Hawk (Falco spar- aged a more rapid southward range expansion, verius), Red-cockaded Woodpecker (Dendroc- Starlings extended their breeding ranges across opos borealis), Brown-headed Nuthatch the entire western half of the continent with ( Sitta pusilla ) , and Summer Tanager ( Piranga its mountain and plain barriers before they rubra), which once bred in the pine lands of bred in the southern half of the Florida southern Florida, on Long Pine Key for ex- peninsula. ample, are no longer found there. In addition, Only seven natural colonizers from the West Robertson has called attention to the scarcity Indies now breed in Florida, and the in- of recent records for the Hairy Woodpecker vasions of two of these, the Cuban Nighthawk (Dendrocopos villosus) in southern Florida (Chordeiles gundlachii, sensu Eisenmann (pers. comm.) and for the White-breasted 1962) and the West Indian race of the Yellow Nuthatch (Sit& carolinmsis) in the entire Warbler ( Dendroica petechiu gundlachi), state ( Robertson 1967). Today apparently have occurred recently. The Florida ranges both are either extremely rare or absent as of five species, the Smooth-billed Ani (Cro- breeders on the peninsula of Florida. Robert- tophuga ani), Gray Kingbird (Tyrannus son (1955: 473) suggested that “since the dominicensis) , Black-whiskered Vireo, and the height of the Wisconsin glaciation the climate nighthawk and warbler, are expanding, in- of peninsular Florida has become progres- dicating that at least some West Indian birds sively more unfavorable for northern species are adapted to the climate of peninsular and that they represent a receding element in Florida. Conditions for colonization of Florida the avifauna.” by tropical birds may be near optimum today. The breeding schedule of the Both the cooler climates of glacial times and (Haliaeetus leucocephalus) and the history of the smaller land areas and larger water barriers the Starling (Sturnus vuZgaris) lend support of interglacial times would reduce the prob- to Robertsons’ first hypothesis. Florida Bald ability of colonization from the West Indies Eagles nest in the winter and migrate north and, in the past, have perhaps eliminated some for the summer, and this has been attributed species that were successful in earlier times. to the summer heat by Van Tyne and Berger If, as seems to be the case, it is the summer (1959). climate that retards colonization of peninsular Starlings were released in New York City Florida by widespread continental forms, one in the early 189Os, and by 1952 their range might expect some of them to be adapting included southern Canada and the entire gradually and thereby extending their ranges United States except southern Florida (Kessel southward. The Red-eyed Vireo seems to be 1953). Even though rural and urban habitats an example. Baynard (1942) considered the had long been abundant in Florida, Starlings species “a regular summer visitor,” but only did not nest on the peninsula until the late “an occasional nester” in Hillsborough River 1930s at which time they were breeding as State Park from 1919 to 1942. In 1967 we far west as Kansas (Johnston 1960). By the found this vireo to be a common breeder in mid 1950s Starlings had bred in Arizona an oak-palm hammock in this park (Wool- (Phillips et al. 1964), New Mexico (Ligon fenden 1967). In 1964 a singing Red-eyed 1961), and Washington (Jewett et al. 1953), Vireo restricted itself to a small area in the and had been recorded from Alaska (Gabriel- mixed oak woods near Sawgrass Lake be- son and Lincoln 1959 ). In peninsular Florida ginning on 5 April. No mate was ever ob- Starlings first bred in Tampa in 1942 (Nichol- served and the bird disappeared after 5 May. son 1964), at Kissimmee in 1946, at in The depauperate condition of the landbird 1956, and at Cocoa in 1961 (Sprunt 1954, with fauna of south coastal and tropical Florida addendum, 1963). The low breeding densities should leave a number of unfilled niches. Of of Starlings in residential St. Petersburg of 2 the 12 species recorded breeding in the Saw- to 6 pairs per hundred acres ( Woolfenden and grass Lake area, only the Yellow-billed Cuckoo Rohwer, unpublished data), as compared is found in closed hardwood forests. No forest with a density of 26 pairs per hundred acres vireo or warbler bred there, although, with in Ithaca, New York (Simmers 1965), and 60 only slight range expansions, certain mem- BREEDING BIRDS OF TWO FLORIDA WOODLANDS 47

bers of both groups might do so. Red-eyed ground in our census areas and they probably Vireos, Parula Warblers, and Yellow-throated had no other avian competitors. Their den- Warblers (Dendroica dominica) breed 15 to sities are much lower in more northern forests, 20 miles north in Pinellas County, and Black- perhaps because of the presence of thrushes whiskered Vireos breed within four miles in and warblers which also feed on or near the mangrove habitat bordering Tampa Bay. forest floor. Twice in 1964 we recorded this tropical vireo SUMMARY in the Sawgrass Lake area; on 14 May two Part I of this paper describes the procedures birds, one of which was singing, were ob- and habitats and gives the results for two served, and on 8 June one singing bird was breeding-bird censuses conducted in Pinellas noted. In Florida the Black-whiskered Vireo County on the Gulf coast of Florida. In- is generally confined to coastal areas, although cluding these, a total of four breeding-bird it is gradually expanding its range inland. In censuses of deciduous forest habitats are now Pinellas County we found a nest near the available for the state. The two additional fresh-water Lake Maggiore in St. Petersburg; censuses were taken in a hammock in the it fledged young on about 27 June 1963. and in an oak-palm hammock in Numerous singing males have been recorded mid-peninsular Hillsborough County. In Part in Brevard and Dade Counties away from II we compare the four Florida censuses with mangrove habitat and as much as 15 to 26 nine taken on or near the Coastal Plain to the miles inland (Robertson 1962). Robertson re- north in Alabama, Georgia, and South Caro- corded 10 and 11 pairs per 100 acres in the two lina. The Florida censuses show significantly years that he censused the south Florida tropi- lower total densities and fewer species. Fur- cal hammock, part of which was a canopied ther comparisons of the species composition forest and which contained no mangroves. In and the fauna1 sizes of the peninsular Florida the West Indies, in addition to inhabiting man- and mainland Coastal Plain areas were made groves, the species lives in forests, occurring using two formulas designed to quantify de- up to 5000 feet above sea level (Robertson, grees of similarity between matched pairs of pers. comm. ). faunas. From these comparisons we con- The abundance of Carolina Wrens and clude that the hardwood forests of Florida Cardinals in the Sawgrass Lake study area have mostly the same species of breeding land may in part be a response to the absence of birds as the hardwood forests of continental true forest passerines. This possibility seems United States, but fewer of them. Further supported by the lower average densities of evidence supporting Robertsons’ hypothesis these two species on 16 censuses of breeding that ’ depauperate breeding landbird birds taken within their ranges in deciduous fauna is the result of present-day climate and forests which did not include pines (see persisting geographical barriers is presented Audubon Field Notes 19621966). On these and discussed. censuses the average density of Carolina Wrens was 17 pairs (range O-42) per 100 ACKNOWLEDGMENTS acres. We recorded 50 and 51 pairs per 100 We wish to thank Olga Lakela and Robert W. acres in the two Sawgrass Lake study plots. Long for their assistance with botanical prob- For these same extralimital censuses the aver- lems encountered in the Sawgrass Lake habi- age density of Cardinals was 35 pairs (range tats, and Richard A. Hutzler who helped us 7-88) per 100 acres; in the Sawgrass Lake make the transects. We greatly appreciate the study plots 40 and 50 pairs per 100 acres were constructive comments of Robert M. Mengel, recorded. In the oak-palm hammock census and the support given us throughout the where four additional forest passerines were project by William L. Jennings and James 0. present, the density of Cardinals was but 27 Bond of the Encephalitis Research Center, pairs per 100 acres. Carolina Wrens, however, Tampa, Florida. The field work was sup- were as abundant here as in the Sawgrass ported by grant no. AI-05504-02 from the Na- Lake study areas, with a density of 51 pairs tional Institute of Allergy and Infectious per 100 acres. Cardinals feed their young Diseases, National Institutes of Health. insects, and it was our impression that these LITERATURE CITED were frequently taken well above the forest BAYNARD, 0. E. 1942. Birds of Hillsborough River floor in our census areas; thus in the oak- State Park. Mimeographed, Publ. by Florida palm hammock the presence of four other Park Service. forest passerines may account for their lower BOND, R. R. 1957. Ecological distribution of breed- ing birds in the upland forests of southern density. Wrens seemed to feed close to the Wisconsin. Ecol. Monogr. 27:351-384. 48 SIEVERT A. ROHWER AND GLEN E. WOOLFENDEN

EISENMANN, E. 1962. Notes on nighthawks of the NICHOLSON, D. J. 1964. Earliest Starling (Sturnus genus Chordeiles in southern middle America, vulgaris) breeding record for Tampa. Florida with a description of a new race of Chordeiles Naturalist 37:91. minor breeding in Panama. Amer. Mus. Nonms, R. A. 1951. Distribution and populations Novitates, no. 2094. of summer birds in southwestern Georgia. Oct. FENNEMAN, N. M. 1938. Physiography of eastern Paper no. 3, Georgia Omithol. Sot. United States. McGraw Hill Book Co., New NORRIS, R. A. 1958. Physiographic and biogeo- York and London. graphic regions of Georgia. pp. 25-76. In T. D. GABRIELSON, I. N., AND F. C. LINCOLN. 1959. The Burleigh, Georgia birds. Univ. of Oklahoma birds of Alaska. Stackpole Co., Harrisburg, Pa. Press, Norman. GRABER, R. R., AND J. W. GRABER. 1963. A com- PHILLIPS, A., J. MARSHALL, AND G. MONSON. 1964. parative study of bird populations in Illinois, The birds of Arizona. Univ. of Arizona Press, 1906-1909 and 1956-1958. Bull. Illinois Nat. Tucson. Hist. Survey. 28:383-528. ROBERTSON, W. B., JR. 1955. An analysis of the HOWELL, A. H. 1932. Florida bird life. Coward- breeding-bird populations of tropical Florida in McCann, New York. relation to the vegetation. Unpublished Ph.D. thesis, Univ. of Illinois, Urbana. JEWETT, S. A., W. P. TAYLOR, W. T. SHAW, AND J. ROBERTSON, W. B., JR. 1962. Regional reports. W. ALDRICH. 1953. Birds of Washington state. Nesting Season. June l-August 15, 1962. Flor- Univ. of Washington Press, Seattle. ida Region. Audubon Field Notes 16:468-472. JOHNSTON, R. F. 1960. Directory to the bird-life ROBERTSON,- W. B., JR. 1967. Regional reports. of Kansas. Univ. Kansas Mus. Nat. Hist., Misc. Winter Season. December l-March 31, 1966- Publ. no. 23. 1967. Florida Region. Audubon Field Notes KALE, H. W., II AND W. L. JENNINGS. 1966. Move- 21:407413. ments of immature Mockingbirds between swamp SUMMERS, R. W. 1965. Residential area [breeding- and residential areas of Pinellas County, Florida. bird census]. Audubon Field Notes 19:622-623, Bird-Banding 37: 113-120. census no. 58. KENDEIGH, S. C. 1944. Measurement of bird pop- SIMPSON. G. G. 1947. Holarctic mammalian faunas ulations. Ecol. Monogr. 14:67-106. and continental relationships during the Cenozoic. KESSEL, B. 1953. Distribution and migration of the Bull. Geol. Sot. Amer. 58:61%688. European Starling in North America. Condor SMALL, J. K. 1933. Manual of the southeastern 55:49-67. flora. Science Press, Lancaster, Pa. KURZ, H., AND R. K. GODFREY. 1962. Trees of SMALL, J. K. 1938. Ferns of the southeastern northern Florida. Univ. of Florida Press, Gaines- states. Science Press, Lancaster, Pa. ville. SPRUNT, A. S., JR. 1954. Florida bird life, with LIGON, J. S. 1961. New Mexico birds and where to addendum, 1963. Coward-McCann, New York. find them. Univ. of New Mexico Press, Albu- VAN TYNE, J., AND A. J. BERGER. 1959. Funda- querque. mentals of ornithology. John Wiley & Sons, New MENGEL, R. M. 1965. The birds of Kentucky. York. Omithol. Monogr. no. 3, American Ornithologists ’ WOOLFENDEN, G. E. (compiler). 1967. Live oak- Union. cabbage palm hammock [breeding-bird census]. MILLER, A. H. 1951. An analysis of the distribu- Audubon Field Notes 21: 635-637. tion of the birds of California. Univ. California Publ. ZoSl. 56: 531-644. Accepted for publication 23 October 1967.