spatial and temporal ecology of oak toads (b l7f0 qi

SPATIAL AND TEMPORAL ECOLOGY OF OAK TOADS (B L7F0 QI'ERCICCS) ON A FLORIDA LANDSC.4PE

As~rri \c r We ~tsecl data frorxr 10 \ ear5 of cortttnitott\, cotlcr~rrerrt 11lttrlitollrrg of o,ik t ~d ( i \ .tt t.lght r\olcitt~d, ephemeral in Florida longleaf p ~ ~ i e - \ ~ ~ t v g r ~ \ s tlplarltli to ,idclrca\\ 1 1: tlrtl :vt.atl~er .iart,il)les affect illrtxement pattern5 of oak toads', (2 i ditX po11d ! t \ d ~ ~ ) l o ~ _"iisl(f tile cotidttto~i of ~1i1r(ii111cI1rlg ~tplnr~ds affect

seleetlon bv adlilt5 or jit\enlle recrttttn~erit~, ,3' \vere p o p r ~ l a t ~ o ~ ~ trends e\ tde~lt j and 131 thct a cla5\tcat ~tetapopul~it~orl rnotlef best represent tllrir p o ' i ) u l ~ t ~ j r ~ t " ~ . ( 1 1 ( ~ ~ ' ~ Of 40'76 o'ik todd\ ~dpt~treitl, 92 2% iaiere adults Sui15tantlal in > 30 ex~ttng jliverl~le\: recr rilt~nent occiil rrtf ordy titree t~trre\ r once eacll at three pond5 dr~ritlg two >ears) Male, oirtnutnhertd fem'iles [alerage for all ve'u-i "2 I i Glo\t capttlres occiit-red (trlr~rlg May-September Adult c'iptures tlrtrtng Jnne-Augitrt tncrea\eei tvith he,t\;ter rdltif'111 htit \\ere not lrlNlie11cecl by tile durations of precedtng ( 3 - f i r 1pt.rlods Moveinent pattenis of ntt.t,urto~-ph\ \uggertect titat oak toads erntgrated when lrloirtrire conctittons t)rco~ne f?\oral-tle Ponti irse bv ,tdult\ w,r\ correlatecl w1t11 rn,h;rlm1tm change In pond depth ( 'tl6i~-Septetinberi fuventle recruttn~ent nrg,ltnel.t. cortelated \mth tniriIrnr1rn poncl depth and the number of weeks 5inee a porld l,ac1c la\t dry, anti pc)\tt~\elv correlated mltll tlie I ~ , I U I Y I I I I ~ niitrtber of week5 ,I pond 11eIcf water contirruouslt The ;iulnlm- of I>rt~ccilng ,diilt\ artel ~t~cc.~rsle rcw utttneiat \yere higheit at ponds wtth~n the 11ardxvooc't-1r1va(Ic~(1 rlpl'tnti i n a t r ~ ~ The tiirectlotl of ino5t : ~ ~ ~ ~ t ~ t g i a t ~ o r i s arid ernlgratlorr\ xva\ nonr,tntlorrl, t)ut rnovetnent occrtrred fro111 ail dlrect~otts allc! the irlezlrr illrrctron of pond entry and eslt did not alw,t\~s correspond 4 total of 91 1% of 1ndi\ ~dtr,ils cvai rec,q?tl~t r x d , 13 3% of hr\t capture\ \v t r~e reeaptlireii ~iurilig the \dtlte \e'11. 'IIIJ 7 7% ~Cttrit~g 'i \ t i b \ e ( j [ ~ ~ ~ ~ t \edr (jrtl\ 1 9% of id1)trli t"ii f)dk t~)rl~I\ t~-ro>xc-d ar-i-iong pond,, mostl\i \mtl~irl a d15tattce of 1 3 h n ifid tlrtt detect d~i~i l t p<)pi~l~~tiOr> tret~d\ abet tile 10- y r studled Presence or ,rl)~erice ,kt ponc.i\ In an> gnen )ear was a poor rntiicator of o t~er~l l (I\<. 11 r \a\v Itttlc evldenee of local evtinrttot> or '>-e\ct~e," but were i~nnhle to detcnntlit~ \vhetller J ~ I V L > I I ~ ~ ~ ~ 2 t'ttlrlltd to rl,itdl p<xlds or cololi~~ed slew pontii for ttrt-ct111ig a5 c~c'fnlti O a k to,tt9 con\~n,i t lor~ can Itest be c>rt\tlred I)\ rnaintainlng rrrulttple pondr withlrl a larrd\cape to incre'tte the prob,tlitltt> of recluttt~tcrlr \~c~ttlttr~ tire lanci- scape neighborhood ilur~rrg at lt-.,tst wtnr vear5 a~ttl at \orme polids at~tf to Increa\e the ltkeltllood of rnter- ~ o n d rnovemetlt

f i y zcortlc Arnphthlntt poprtlatton\, 4nurar1 itreectsl-ig, Bufii (pr(. t i 115 E ~ ! I ~ I I I ~ T ~ polid5 Nvdloperiod. kfetdpopulatton, Oak tohds, Ter11po1'wy ~~etl,urcl\, ?I etlancl\

A DEARTH of long-term studies of amphibian populatio~ls restricts our uiiderstanding of population dynamics and our ability to distinguish true population trends from natural fluctuations (Blaustein, 1994. The few long- term studies typically have been conducted at a single site (Dodd, 1992; Senilitsch et al.. 1996). Such studies provide important infor- mation about life history and patterns of pond use, b u t cannot probide the larger, landscape- level perspective that is gained by sampling multiple ponds within a landscape. The logis- tical and fiscal constraints that prevent intensive sampling at multiple ponds often lead to "snapshot" sampling to assess amplribian distribution or richness by presence/absmce

CORR~~POLI )F \ ( F e-rndt1. kg~eenl?rr"fs fed tti

s n ~ ~ e v s iSkellv et al.. 1999: Skell! et al., 20031. Such \urve js are likely to mi\s species presence and certainly i~liss informatiot~ on population and metapopulation dynamics.

Metapopuiation theorj. which holds that subpopulations becorne locally ext i~~ct and are recolonized h\: ~nigrations from other sub- poptdations i Hanski ancl Sinlbrrloff, 19971, is a popular paradig111 and is increasingl) ap- pliecl to amphibian ecclloa. hfost studies use presenceiahscnce sun cys at ponds to determine the occurretice of yopnlations. Presence of a species foilc~wed by absence (no detection) the followring ?ear indicates extinction, whereas abse11cc h l l o ~ ~ ~ e d by presence during the next year indicate recctlonization (see review in Marsh and Trenharn, 2001). This "ponds as patches" (Marsh and Trenhan~, 2001) approach ;issumes. in part. that (1) s n w e ~ ~ s at

polids detect the presence or abse~ice of a specie": (2) absence is in i'act an indicator of local estinction; and ( 3 ) recolonization. or inter-posid moven~ent. is a cornlmn occurrence l ~ u t decreases witll increasing distances a~nong ponds. However, each of these assunip- tioris is questionable. Man)- amphibian species skip breeding during sonic. years but are nonetheless present (i.e., not evtitict) in the surrounding np la~~ds (Semlitsch, 1996). Man!- amphibians also visit ponds mailily during short, explosive breedilig periods, increasing the likelihood that they will be considered "absent" (Greenberg and Tanner, 2004). Finally, many pond-breeding amphihiall species are highly philopatric, thus reducing the likelihood of inter-pod movement (Blaustein et al., 1994). The temporal and spatial variability in breeding, recruitment, population structure, and metapopulation organiza- tion-or lack thereof-can be captured only with intensive, concurrent, and continuous mark-recapture sanipling of multiple ponds within a landscape over an extended period.

Like rlialry amplribian species in tlie soutli- eastern United States, oak toads (Bz~fc) pet-ci- cus) breed priimarily in isolated, temporaly, fish-free ponds (Moles and Franz, 1987) but inhabit tlie surrounding uplands during most of their adult lives. Hamilton (1955) described them as "extraordinarily abundant." Carr (1940) wrote that . . . "Today I believe it would be possible to cross central Florida by car and at no time be iri a situatioii where tlie calls of qnerciczis were not audible." Documentation of their basic siatural histosy has been limited primarily to obsesvation {Carr, 1940; Hamil- ton, 19Eifj; \Ysight, 2002) or studies at a single site (Dodd, 1994). Because there have been no studies of oak toads at a broad spatial and temporal scale, little is known of how the condition of s~irroundiiig uplands affects poiid use, llow weather affects activity and l-trccding, what factors influence the selection of hreed- ing ponds by adults, <)I- the likelihood of successf~tf recruitinent. To our knowledge, no studies liave addressed the cpatial and temporal dynamics of breeding ecology and recruitment. or wliether a metapopidation model best describes the populatioii ecolo{g of oak toads.

\Ve used data fi-oin 10 ?-ears of continuous, concurrent monitoring at eight isolated,

ephemeral ponds it1 Florida longleaf pine (Pirlrrf pcrZrrstj+s)-\\<I-egrass iilristickr beyt-ichi- ma9 uplands to c~amine spatial and te~nporal ecology of oak toacls. Half of the ponds were in sandhills that lratl beeti itivaclcd by harct\4foods and sand pine (Pinus clarrsa) as a consequence of fire suppression, and half were in regularly burned, savans~a-like sandhills. \lie asked (1) did weather variables affect movement patterns of oak toads?; (2) did potid bdrology and the condition of surrotinding uplands affect pond selection by adults, or juvenile recruit- niesit?; (3) were population trends detected?: and (4) did a classical metapopulation s~iodel best represent the spatial and temporal dynamics of oak toad breeding, recruitment, and movement?

Our study ponds were eight small (0.1-0.37 ha), isolated, ephemeral sirtkhole ponds i11 longleaf pine-wiregrass sandhills on the Ocala National Forest, Marion and Putnaln Coun- ties, Florida. Ponds 1-4 were in sandhills that had been invaded by hardwoods and sand pine as a result of fire suppressio~i. Despite at- tempts at habitat restoration (hardwood reduc- tion) by burning at 1-4 year intervals during the past two decades, hardwood density and wiregrass cover remained patchy. Ponds 5-8 were in savanna-like sandhills with a continuous wiregrass ground cover, widely spaced longleaf pine trees, and few hardwoods, Since 1976 this uplalid matrix has been burned regularly at 2-3 year i~itcsvals (for more de- tailed habitat characteristics see Creenberg, 2001). Ponds 1-3 were closely spaced (all within 132 m of one another) within 10-30 111

of the regularly burned sandliills on one side, sepal-atecl by a sand road. Ponds 1-7 were all within about 0.7 km of one another. Ponds 7 and 8 were appi-ouimately 9.5 kin south of the otlrers.

\"C7e installed dl-ifi fknces 7.6-n~ long and spaced 7.6-111 apart around the perimeter of each pond near the high kvater line, such that 50% of eacli pond was fenced, with fences and spaceseequallj distributed arouiid the entire pond. The distances between fences and pond eclges var-ied wit11 pond depths. We placed

pitfall trap" 19-1 buckets) inside and outside of each end of each fence (four l)uc.kets per j;iicel to detect difectional nlovement by aniphibians to and frola1 ponds. 'll'e placed a sponge in each pitfatt trap, and moistened them as needed during trap checks to reduce the likelihood of desiccation. We positioned a double- or single-ended frtnnel trap at the midpoint of each fence on both sidcs (tm7o per fence).

147e checked the traps three times weekly fiom February 1994 through January 2004. We n~easured the snout-vent length (SVL), and mass to tlie nearest 0.1 g using a hand-held Pesola spring scale, of all juveniles and about half- of all first-capturctl atlults (sonle were not measured if capture rates were liigli). Adults were sexed and marked by pond number and year of capture by toe clipping. Males coulcl be distinguished frorn &males by having a dark- colored throat and dusky venter (FVright, 2002). Oak toad lnetaniorphs could be distinguished from southern toad ( B . terrestm's) rnetamorphs by their more prominent light middorsal stripe, often orange toes, and head shape (southern toads have a Inore spatulate head and flatter nose). Newly metamorpl-losed toads were not marked, as they were too small to toe-clip (mean SVL of juveniles June- October, all years, was 11.7 1x1117 but many were 6-8 mm SVL). Animals were released on the opposite side of the fence frorn their capture point.

We measured water depths weekly beginning in March 1994, and recorded temperature and rainfall daily at 0700 11 hegil~ning in April 1994. Beginning in January 1997, we recorded minimuin and n~axilnum daily t e~n- perature, and barometric pressure and relative humidity were recorded using automatic lo@ ers at 3-11 intel~als. \Ve used a Geographic 4 Positioning Systern IGPS) to estiniate inter- pond distances.

Statistical Annlytses

We risecl 1L'right's (2002) caliort classifica- tions based 011 estimated groxvth rates to classif>- iuctividuds as < 1-yr old 'weniles" (<20 unmi, or l-y--old (20-2'3 mm), or 2-yr-oltl (30-39 1m1i1) "'adults". Both the juvenile and 1-yr-old categories likely included sonae non- llreeding sribadults. For statistical analyses ts7e considered all incli\idnals 220 mi11 as adlilts.

\f7e used repeated rneasnres i ? - t z 7 ? Analysis of Variance jA4~ol7~i to test tcjr differences in pond 11~e hett\7eeii upland matrices and anzortg \-ears all adults \etiterir,t: and exiting), and b:, recrriits insing onl! exiting neonates as a consen7ative estimate of recrnitiizent). Data were standat-dized to nurnl~er of c:aptures per 100 trapliigl-lts fot- tllese analyses, because ponds of d ik r en t sizes had clifferelit numbers of traps. and some traps were floodecl (close4 during high water. Virtt~all) a11 (98.3%) movement by adult oak toads occrlrred during May- Septeixher, arid niost (95.0%) movement by juveiiiles occurrecl during June-October. We included only tlrcse rnontlas in our analyses o f ~~plarid n~atrix treatment effects to reduce potential bias i~~troclucecl by !-laving inore traps closed and clifferent proportions of total traps closed arnong ponds dl~riiig winter flooding, when oak toad captures were negligible.

We used stepwise multiple regression to explore the relationship between landscape- level (ponds cornl~ined) movement (captures per 2-3 day trapcheck interval) of adults and jrlveniles (separately) and weather variables including: cumulative rainfall, number of prior 2-:3-&y trapclieck intervals with no (<0.3 cm) rainkll, teunperature (mean minimum, minimum, mean maxinrurn, inauiinuin. and mean daily), barometric pressure (mean daily, minimum, i~axilntxm, and mauimuim change per 2-3 da): trapclieck interval), relative humidity (niean daily, minimum, inaximnm, and maxi- inr~ni change per 2-3 day tl-apcheck interval), and an interaction terrn between cumulative precipitation ancf rnaxirntlnl change in barometric pressure per 243 day trapcheck interval. For kidnlts we performed two stepwise regressions: one using both year-round data, to determine \vl~ether weather variables affected tlle cornrnencenrent of activity in spring, and another using only J~1ne-~4ugust data (\vlieti 92.6% of captures occurred), to examine tlie relationship between wcatlier variables and capture rates drlriiig their peak active season. For jnveniles we limjtetl our correlation data- set to months wlien the> were i~ictst frequently captured (Julie-Octoheri artti years when recrrxitme~it was slil)stantial ( r l > 30) - 1994, 1998, and 2002.

To assess bv11c.thc.r hydrolog!- iiffected pond selection hy bl-eediq ;u.lults, we used step.isise ~nriltiple regression to correlate rainhll and

hydrolz~gical c h a r a t e - i s of ponds dt~rilig _Zf~l~-Septenhet-, atid the number of first- captured adults :"La\-Septer~lber). Capture data ~\~c.re sta~rcfardizctl to number per 100 trapnigllts anti log-tra~isiitr~~~ecl. Independent variables (per pond for 10 \-c.;lrs: ,a = 80) were: riun?ber of m-ceks a pond wils dl?.: number of weeks a pond was xx7et ( 2 2 cr-rtj: minirnunl, ~-r~;~xir~it~ni, and maxinlum change in depth; nrrmber of times a pond refilleel atier d e n g ; ~lumbel- of iveeks since a pot~d wis last dr). (0 if it was dr)- tlr~ri~ig the f is t m7eeIc of Ma>; when the data period began); rnaxiliii~m number of weeks ;I poncl continuously lleld water; total precipitation; and iin interaction behveen the maximurn number of weeks a pond continuously Iield water and tlie number of weeks since a poncl was last dry.

To assess whether pond hydl-olocgy ancl hreeclir~g effort affected recrriitrnent rates, we used stepwise mi~ltiple regression to correlate the nilrnlm- of breeding adults, rainfall, and hydrological characteristics of ponds (using the variables described above) during Ma>-Septei-riber with oak toad recruitment (exiting jrlveniles, June-October). We used st;lndardized (per 100 trapnights) capture data fbr adults :md recruits.

We useel a compass to obtain an azimuth from the center of each pond to every pitfall trap. Undcr the assilmption that captureshy intermittent, eqlrally spaced clrifi- fences permitted detection of directional movement i l ~ some (if riot 50%) proportion to its occurrence, we used the nonparalnetric Rayleigh test (Zar, 1984) to detei-nliile wl~etller tlle annual emigration or immigration ~~~overnents of adults or juveniles (if )I 2 30) were nonrandom.

We used paired t-tests to coinpare the proportion of first-captured males to fem;iles per pond, and a I-way ANOVA to examine wliether tlie proportion of' males cliffired ;imong years, usisrg ponds as replicates. One- \I-* .4NO17A also was useel to assess whether the propcxtic~n of 1ni'lesto females tlifferecl amorrg ~no~~ t l i s , using years as replicates (ponds cornbinecl). Only months (for analyses using mo~ithsl, or years (for analyses using \-e;trs) 1i;nirig 12 2 10 se.iet1 adults were included in allal? ses. Stuclent's t-tests were tised to cort~pare SVL and mass between male at~cl fertzale oak toatls. We or~iitted 1994 data frorl-t these anal) ses because oak toads itrere

not sexed that year. I17e cxi\rr?ined adult population~ trends 01 cr the LO-j ear stud!, pel-iorl using Pearson's product-ii~oment correl;-ltions behvccn st-ct+~ year ancl number of first- captured adults. ?Ve used scluare-root or log-transformations on data to redwe hetero- scedasticity. All proportions n7ere arcsine square-1-oot transformed fbr anal? ses.

P(>prr/crtion Sttrrctu re

?Vc captrtred a total of 3076 oak toads (luring our 10-yr study. Most (92.2%) were adults (220 mm): relatively k w juveniles (7.8%) were captured. Snbstalitial recrui t~~le~l t ( a 2 30 juveniles exiting a pond) was detected only 3 ti~nes (at Ponds 2 and 4 during 1994, and at Pond 3 in 1998). Males (69.4% ri: 2.1%) were captured inore frequently than females (average for all years 30.6%~ t 2.1%) during all years (t = -3.61; df = 170; P < 0.0001). However, the proportion of males was significantly higher in 2001 and 2002 than in most other years (there were fewer males in 1998 tlian in 1997. and 1997 die1 not differ fi-om 2002) iFh,?z3 = 12.98; P < 0.0001). The proportion of rnales to females was sirnil:ir among tested montl~s (June-September) iF : 29 = 0.66; P = 0.5855). Adult males were srr~aller (mean SVL t SE 28.7 t 0.1 mm; t = 15.81; cif = 1388; P < 0.0001) and lighter (rnean ~nass t SE 2.1 t O.0g; t = 17.70; df =1313; P < 0.0001) than females (mean SVL t SE 31.5 rt 0.2 mm; Imean illass t SE 2.9 t 0.1). Mean ( t SE) SVL and bociy mass of juveniles (defined as <20 mm) were 11.8 t 0.2 lnln (range 6-19 mm) and 0.2fj t 02 g, respectively.

Oak toad activity was highly seasonal. Nearly all (98.5%) adult captures occurred cluring May-September; peak move~~lent 192.6%) ivas June-August (Figs, 1. 2). Most (95.0%) mover~~ent by juveniles occltrred durirtg June-October. ??'hen year-rord data were used. mox7enlent by adult oak toads was positively correlatecf with c11rntll;iti\7e rainfall iP < 0.0001), rnaiirnurn temperature (P < 0.0001), and an interaction between curnrxlative rainfiall per trapcheck inteival ancl the number of prior trapcheck intends having no rainf"tll (P < 0.0001) (Z;;Hl,, = 99.66; P < 0.0001:

326 EIERPETOLOGICA [lid. 61, A70 4

I j m ~ d u t t (n = 3759) i

1 2 3 4 5 6 7 8 9 1 0 1 1 1 2

Month

FIG 1 -Montllly proportion (all years cornh~ned) of clcl~rlt and juventle oak toad\ captured during Feb 1993- Jan 2003 at elght isolated, ephemeral ponds w t h ~ n longleaf plne-mregrass simdh~lls, Ocala RJat~onal Forest, Manon and Putnarn Count~es, Flonda, U S A

r2 = 0.27; RMSE = 1.31). Movement of adults during their peak active season (June-August) was positively correlated with cumulative rainfall (P < 0.0001) and negatively correlated with inaxiinum barometric pressure (P = 0.0177), iF2,2.in = 108.82: P C 0.0001: r" 0.47; RMSE = 1.82) (Fig. 3). Movement by juveniles (June-October) was positively correlated with maximuln temperature ( P < 0.0001 ), mean daily barometric pressure (P = 0.0094), and an interaction between cumulative precipitation and the number of prior trapcheck intervals having no rain (P < 0.0001), and was negatively correlated witlt tnawimum change in relative humidity (P = 0.0396), (F4,]2* = 24.53; P < 0.0001; r2 = 0.44; RMSE = 0.51).

Infience of Hydmlogy of2 Breeding and Recruitment

Pond use by adult oak toads during May- September was correlated only with 1navi111u111 change in pond depth (P < 0.0013) (F,,;; = 11.08; P < 0.0013; 3 = 0.13; RMSE = 0.99). Oak toad recruitment was negatively correlated with ntirrimum pond depth ( P = 0.0005) and the number of weeks since a pond was last dl?; (P = 0.0024)) and positively correlated with the mzimum number of weeks n pond continuously held water (P = 0.0001) iF3 .;. = 11.84; P < 0.0001; r' = 0.27; RMSE = 0.34).

The number of breeding adults at ponds xvas higher within the hardwood-invaded upland

matrix than in the savanna-like longleaf pine- cviregrass matrix (df = 1; F = 8.26; P = 0.0058). There was a repeated rneasures e&ct (F = 3.44; df = 6: P = 0.0010), but no effect of year, nor was a year X treatment interaction detected (P 2 0.23). Similarly, significantly inore juveniles were produced from ponds within the hardwood-invaded upland matrix than from ponds in the savanna-like sandhills inatrix (F = 25.31; df = I ; P < 0.0001). Recruitlnent also differed among years ( F = 6.63; df = 9; P < 0.0001). There were more recruits in 1994 than in other years, and the number of recruits in 2002 did not differ from those in other years. A year X treatment effect was detected (F = 6.72; df = 9; P < 0.0001), but there was no repeated rneasures effect (F = 1.20; df = 6; P = 0.3184).

Orientation

Many (27 of 40) of the adult irnrnigration and emigration movements, and all t l~ree emigration movements by juveniles were significantly nonrandom (P < 0.05) (Table 1). However, many oak toads immigrated to and depai-ted from ponds in all directions. Further, the mean angle of adult emigrations often did not correspond with the general vicinity of mean immigrations at the same ponds and during the same years. For example, in 1994 at Ponds 1 and 4, and in 1999 and 2002 at Pond 3, adults entered and exited in a nearly straight-line direction (rather exiting toward the vicinity of entry). On other occasions-for example, at Pond 2 in 1994 and 2000, and at Pond 6 in 1998-most oak toads emigrated in the same vicinity where most immigration had occurred. However, because oak toads were not marked individually we could not track lnovements of indi\~duals. Directions of rnove- ments were infrequently tested at the same pond(s) during multiple years. At Pond 2, the mean angle of both irnrnigration and emigration rnovernents was similar during 1994 and 2000 (entering from and exiting toward the northern portion of the pond). At Pond 3 individuals entered froin the west-southwest (moving east-northeast) during 1999, 2002, and 2003, but iminigrations were not signifi- cantlj- directional during other years (1995, 1996, 1998). Einigrations were generally east- southeasterly at Pond 3 during 1996, 1998, 1999. and 2002. Adults entered and exited

200 Pond 1 1 1 1 rS Pond 5

200 Pond 4 (0.37 ha)

1994 1935 1395 19.37 1938 1999 2000 2301 2COi 20C3 1594 3595 1996 19~~71955 199cjLCli;O 2001 200'1 .?CE

YEAR YEAR

-

FIG 2 -Total first captures of oak to,td actrllts :;Ma>-Augn5t) and rrcnnti iex~t~ng lu\en~les) (Jtme-October) and cumulative ra~nf'dl and mdxlmuni pond riepth dunng 'Lla\--\ugu5t, 1994-2003 'tt elgilt polrds In the Ocdld Nationcl1 Forest, Marlon and Putnanr countlei, Florrda C S A

- ~ d u ~ t (2 20 mm SVL) Pond 8 Juvenile(<2OmmSVL) (0.35ha)

A Rainfall (cm) U M a x i m u m D e p t h ( c m ) 8.

p A /

J

Z - 9 zoo t-

Year

F I ~ , 3 -Kcldt~oilsl~~p bch\t1t'ii tt~ttr (stlttly \tlclr) ditd rt11tnbt.r of ,id~ilt (220 rrrrn SVI,) o,lk toacls first-t,iptrired at elglrt 1so1,itetI c~phc~~ner,il pontls Oc,ila Nclttonal Forest, M,tnon ~ ~ n t l Ptltnarri (;o~tritlrs. Flortda, U S h Nilnrber\ of .~dtilts prt.\e~itc~tf hrre are iitltl~i~~\fOri~ieii, 1)ut wele log- tran\forr~ncd ihr eorrelat~oli

recaptrise rates also occurrecl in 1999, when 14.5% of individltals t l~at were first captured it1 1998 were recaptlisecl. D-rsring 1998 we recaptured 10 (4.7%) individtrals that were first capturecl in 1995. These recaptured oak toads were at least 3- and possibly 4 years old, since we marked only large juveniles (> about 15 rnm SVL) and (primarily) atlults that pl-obably were 1 year old in 1995. h total of 19.1% of oak toads was recaptrrrecl

at the same pond as their 01-iginal capture (Table 3 ) . Only 1.9% of captured oak toads ~nroved anlong ponds: nrost exchange was behveen Ponds 1-3, all within 132 rn of one anotl-ier. One individual that was originally

captured at Pond 5 m7as recaptured at Pond 6. about 369 m away. Because most juveniles were not ixarked we were unable to determine whether they retrirned to natal ponds for breeding as adr~lts 01- colonized newr ponds.

At least one oak toad was captured each year at each pond with fe~v exceptions (Pond 5 in 1996; Ponds 3, 4, and 5 in 1997; Pond 7 in 2000; Pond 4 in 2003) (Fig. 2). Presence did not necessarily indicate ovirall pond use. For e~arnple, although oak toads were present at Porid 7 driring all years except one. they visited in sllhstantial n~~tnbers only during 1996 and 2003. Siinilarly, absence from a pond probably did not indicate local extinction. For example, oak toads were absent fro111 Pond 3 driring 1997 (a ctry year when few oak toads were captured at any pond). but bred tlrere in high numbers there during 1998. Nearly all breeding adults were first-captures or recaptured at the same pond where they had been captured during a prior year, which indicates that they did not "rescrle" Pond 3 froin other ponds, but rather skipped the 1997 breeding season and remained iri the surrounding uplands.

Oiir 10-y1- study illustrates the variability of oak toad breeding and recruitment among ponds and years. A study that included only one or a subset of our study ponds or years might have produced different results. For example, several ponds were little-used for se~7eral consecutive years te.g., Ponds 1 and 2, 1995-1999; Pond 7, except in 1996 and 2003;

T t e ~ t. 2 -Total ni~trtbrr of \,ttne-~~e,tr i S l ) aricf 111iilt1-yedl (MY) adtilt o,ik toads rec,ipti~i-etf for 10 vr (Fel.). 1994-Jan. 2004) at elgilt tsolatcd epl.letnen~1 poncis, Ocal,~ Na t~on~~ l Fore\t, Mario11 ,mtl Piitnarrl courrtte5, Flortda. U.S A

Vt,ctr of bt \ t c,tptrirc2 tic c dpt111~ l o t 1 Tot I'elcrilt

i Ket .qI" \ ( '11 I 3 1 i 3 6 7 X 9 I 0 ( .ip W'"C.LP wceq>

Tot Hcp Tot. '2 SY Hec,cp Tot % MY Hecap

T\iii,t- ,3 -Total t~~tmber and percentagr of adult oak toaci5 rec,lptrtrtd 'it tlir i,unc. iSPi or ,irlotiier potltl il l ' as hr\t tdptrtre C ~ I I ~ I I I ~ 10 \ r (Frh 1993-Ian 20011 ,tt e ~ g l ~ t irolatt~tl ephemeral pond5 Ocala \ a t i o~ l~~ l Fore\t, X f a t - I ~ I I ar-icf

Prrtn,itn conntit.5. Florrd,i II S .1

Tot Tot. % SP Recap Tot. 5% IP Recap

Pond 5, 1994-1997 and 2001-2003). Others (eg., Pond 6) were used by adults on a I-eg~ilar basis but apparently never producled substantial numbers of juvenile i-ecn~its. Data froin only these ponds or years might have been interpreted as showing that oak toad populations were declining, whereas our long- term landscape-level data suggest otherwise (see Blaustein, 1994).

The relative landscape-level stability of the oak toad population was incongruous with the low capture rates of juveniles duriilg ixost years and at most ponds, especially because only minimal short-distance inter-pond ixove- merit occurred. Highly directional movement patterns by juveniles, oriented exclusively within the 7.6 rn spacing between fences, could have resulted in our not detecting recruitment events, but seems unlikely. Fur- ther, we did not obseive oak toads climbing over fences (e.g., Dodd, 1991) or climbing from pitfiall traps. However, on one occasion juveniles were observed avoiding traps by moving around the rims of pitfall traps (Ray Ashton, personal obsei~~ations at Pond 3, 2003). Yet, capture rates of i-netamoi-phs of the closely-related sonthen? toad, \vhicli are similar in size. were mucln higlie1- ka . 1000 captures) during soine years, suggesting that capture rates also sliould reflect the rebtive recruitment rates of oak toad metanlorplis. Confusion between oak toad and southern toad rnetaiiiorphs was likely minimal, 21s metamorphs of these species h x e distinguish- ing characteristics. Also, most southerii toad recruitment occr~rrcd during April and itlay

(with less in June and July) whereas most oak toad recruitment occurred during June-Octo- ber. If bias by trap-avoidance occurred, we assume that it was consistent among years and poilds, and therefore comparisons of recruitment among years and poncis should be valid if used as an index.

In our study, some ponds (e.g., Ponds 5 and 6) apparently produced virt~ially no juveniles, yet maintained moderate to high capture rates of adults dui-ing niany years. Dodd (1994) sampled an ephemeral sandhill pond for 5 years usiiig pitfall traps with a drift feiice that completely encircled it. He also reported relatively high adult capture rates through the third of five breediiig seasons sampled, even thougl~ the pond held water for more than a few weeks only during the first (1985) aiid third (1987) breeding seasons, and no juveniles were produced. Despite this, Dodd 11994) captui-ed juveniles (<23 mrn SVL) entering and exiting the pond duriilg fall 1988-sumi~ier 1989. Senilitsch et al. (1996) reported only one or two years of substantial recruitment at Rainbow Bay iii South Carolina hi- inost anurans during a 16-yr period. They suggested that population persistence at the Bay was aided by 'I-escues" from otlier ponds. Ho\vever. we obsemed a pattern of coiisis- t e n t l ~ ~ low recmit~nent at all eight of our stud>- ponds. reiidering this explanation less likely.

Sinscli (1992) found that inetaixorphic nattei-jack toads iB. cnlnt,~ita) emigrated fi-om iiatal lakes or ponds but shou~ed high breeding site fidelitl, in subsequent years. Becatise most juveniles m7ere not ~~narked we could not assess

whether they returned to natal ponds for hr-eeding as adults, or colotiized new ponds. However, we captured very few jut.i.ni1es enteling study ponds? indicating that the?. were not immigrating to nonnatal ponds, at Least as jul-cniles. Stuclies show adults of many amphibian species are philopatric (Rlaustein et al., 1994i: whereas others readily colonize nebvlj--created pctnds (Petranka et al., 2003).

Our data suggest that most (>93%) oak toads breed during only one season. During his 5-year study, Dodd (1994) also reported that rnost recaptures occurred during the same season as their first-captnl-e. Hamilton (1955) surmised that male oak toads first breed drxiing the breeding season following transfornlation, and females during the second. This could explain why we captured (on average) more than twice as many males as fernales each year. Dodd (1994) also found that males outnum- bered females by 1.75:l at a north-central Florida pond.

Our multi-year recapture data suggest that mortality was high during most years, but apparently was reduced during years when breeding was s~lspended (e.g., 1997). Haniil- ton (19517) estiinated that oak toads live 2-3 years. Assuming that oak toads first breed at age 1, our data also indicate that inost live for about 2 years. However, some (about 6.4%) lived longer. Some individuals apparently lived for at least 3, and possibly 3 years. We may have underestirnated age at first breeding if breeding was suspended during some years. Similarly, our longevity estimates could be wrong if age at first breeding was >I year, or if oak toad age cannot be estimated accurately by size due to deterlninate growth or slowing of growth with age. The possibility that oak toads avoided recapture by trap avoidance renders our conclusio11s ahout age somewhat speculative.

Oak toad activity mas governed primarily by season. Most (98.5%) adult captures occurred durirrs Mq-Septenibcr, and most juveniles were captured during June-Octobel-. Dodd ( 1994) reported similar annual actik-ig- pat- terns for oak toads. Our results showed that mi~ximrrm temperature, crxrnulative rainfall, and rainfall patterns were correlated with \-ear-round adult activity pattei-ns. Clearly, seaso~ial cues such as temperature or photo- period (\vllich are autocorrelated) play a role in

annual activity patterns of oak toad?. Oak toad actixig- is comrnonl! associated with heav? rainfall (Carr, 1930: Hai~ilton, 195Fj; Wright, 2002). Dodd ( 1994) reported that rainfall influenced movement, especially if a heavy rain was preceded by a dr>l period durirrg summer; if rain fell regularly, its influence on ~mvelment \tias not pronounced. In our studv. variability in adult capture rates was better explained (higher ?) by weather variables w11c.n only peak breeding season data (June-August, when 92.6% of adults were captured) were used. During peak breeding season, rainfall and barometric pressrrre had the greatest influence (arnong tested variables: 011 adult oak toad activity. In contrast, juvenile captnre rates were best predicted by niaxiinurn temperature, mean daily baroinetric pressure, and a11 interactionl between cumulative precipitation and the number of' prior trapcheck intervals having no rain. Our results indicate that adults and jrtveniles may respond differ- ently to rainfiall patterns. The nunlber of June- August adrllt captures increased wit11 heavier rainfall but was not influenced by the length of the dsy period preceding rain. In contrast, metamorphs may remain near their natal ponds until moisture conditions itre favorable for emigration to the surrounding uplands.

Our data suggest that pond hydrology- especially hydroperiod, water level fluctuations, and pond depth-influenced use by adult oak toads, but pond use varied considerably regardless of those hydrological characteristics. For example, adult capture rates were high at Pond 3 in 1999, althougl~ the pond was dsy during the entire breeding season. Simi- larly, the numbers of oak toads using ponds and annual patterns of pond use varied considerably among Ponds 1, 2, 4, 5, and 6, although all had sinlilar hydrological characteristics (depth and hydroperiod). Dodd (1994) suggested that fewer adult oak toads were captured at a north-central Florida pond after 1988 because they moved to other nearby xvet1ands to breed. Our capture rates at all ponds also were negligible when May-Sep- ternbcr precipitation was lo\v (62 cm) in 1997, as were water levels (maximum depth 36- 63 cm at wet ponds; 2 ponds were d n ) . How- ever, we found that inter-pond movernent by adults was infi-equent and inostly short- distance. Our data suggest that oak toads

suspended breeding acthi%, rather than n-ioking to another pond, during drought.

Several hyclrological characteristics were correlated with oak toad recrt~itr~tent. In- creased recruitnient was associated with longer hydroperiods, probably because longer liydroperiods gave tadpoles adequate tirne to corliplete transformation (Pechrnartn et al., 1989). Breeding occurred over a period of several months, tlrr~s an extended lrydroperiod during a11d beyond tlie breeditig period would all on^ 12io1-e tadpoles to complete transformation, which takes about five weeks for oak toads Wright, 2002). Similarly, ~minirnu~ii pond depth ctllri~lg the oak toad breeding season- an indicator of pond dry-downs and tadpole desiccation-was negatively associated with oak toad recruittnent. In contrast. pond d ~ y - downs immediately before the breeding period had a positive infiilence on recruitnient. This may have resl~lted from reductions of invertebrate predators im~nediately prior to the appearance and development of tadpoles (e.g., Scmlitsclr et al., 1996).

Breeding effort (adult first-captures) was not a significant predictor of oak toad recruitment in our nlodel. As further evide~ice that breeding effort does not necessarily correspond with recruitment levels, we found only four tadpoles at only one of the eight ponds (Pond 2, 26 September 2001) during 2000-2001 box sampling for tadpoles (sampled at 3-4 week intervals at all eight ponds, if they were wet), despite moderate adult capture rates at most ponds during both years (A. Storfer, C.H. Greenberg, and G.W. Tanner, unpublished data). Clearly, successful recruit- n-ient depends on a complex interplay between adult breeding effort, competition, predation, and edaphic betors such as weather and pond hydrology (Alford, 1989; Petranka and Ken- nedy, 1999; Wilbur, 198'7).

We found that both 11l-tl~ibers of breeding adults and recruitment levels were higher at ponds within the hat-dwood-invadcd sandhills than in tlie savanna-like sandhills. However, this should be interpreted wit11 caution. A repeated nleasures effect suggested that site fidelity by oak toiids cor~ld affect which ponds are consistently l-lsed most heal~ily. Further, habitat patchiness within the harcl\vood-invaded upla~id matrices, and the prtjximiv of tl?p h5'C) treatments, could co~nplicate ititer-

pretation of' these results. However. given the relatively wide distrihutioti of oak toads withi11 pine flatwoocls ant1 xeric habitat (Carr, 1940; T\;t7rig11t, 2002), it is not s u ~ r i s i n g that oak toads are to1er:lnt o f habitat changes associated with fire suppression. at least in the short tern^.

I13 southeaster11 fire-r~~aintained habitats, many aniphibim species nt;~!; tolerate harcl- wood invasion (Greenberg. 199:3; Greenberg. 2002; Litt et ill., 2001; Meshaka and Layne, 2002; Palis 1998). Greenkerg and Sirnoirs ( 1999) suggested that topographic and climatic variability, as well as spatial patchiness and variability in fire fi-equency, season, and in- tensity, liistorically permitted oaks to reach tree size in varying densities over tirne and across the high-pine landscape. Amphibians that evolved in a spatially and temporally dynamic landscape are likely to tolerate transient couditions within tlie upland matrix.

Although many acj~llt movements were 5ignific;mtly t1irectiori;-~l, otliers i32.,5%) were not. Many ii1divid11als i~n~nigrated and emigrated from all clirectio~is even during move- inents that were statistically nonrandom. Dodd (1994) reported that immigratirzg females, emigrating juverriles, and males snoved, on average, in nonrandom directions, but that there was a great deal of variability in orientatiot? within tliese groups. Because we did not give oak toads incli~ic111al markings, we could not cletermil-ie whether individuals returned to their point of origin. However, within-season immigrations and emigrations at any given pond often were in different directions, suggesting that oak toads did not always return to tlre vicinity of their origin, or that oak toads from the entire surrounding upland neighborhood rlsed p o d s during a given breeding season, or both.

Directional movements are not necessarily indicative of upland habitat tise. artd oak toad directional rrnoverne~~ts did not appear to be particularl! associated wit11 either the harclxliood-invadecl or sl-~vvanna-like upland matrices surroundiitg or adjacent to pontfs. For example, adults apparentlj entered Pond 1 fro111 liarclxx7ood-i-t~vacfedle sanclhills, hut euitecl to\\-ard the adjacent sktndy road and savanna- like sandhills. In contrast, Dodd (1994) reported that oak toads tendecl to enter liis study p d froln and exit tow;trcl a santlhill cotnmu- nity to the wcst-so~tthxv-ect, rather tban the xeric

lialriln(,cks to the north. Dodd (1996) cap- and ponds. and retain the possihili? of inter- trlred oak timls rip to 911 n~ fro1l-i the near- pond mo\-rr~~e~it . est r+etland.

We h r t r l c l little e\iclerice of a inetapopu!a- tion strilctul-e of oak toads within cmr s t ~ d y area. We had i.\v i~istances of apparent "extinction." Absence froin a pond during ;i given ! eai- nlas often fbllon7ed b!; breeding in lrigh nrinlbers at that pond dr~rirrg the next year. Senllitscli et a1 (1996) suggested that migration ii-oilr nr;irhy ponds explained tlie persistence of' breeding sonthern toads and spadefcmt to:ids ( I i s holl~mokii ) at Rainhow Bay, despite infi-equent juvenile recruitn~ent. In our stirtly only a few adults ~tioved atnons poncls, ancl tile11 largely only among 1-elritively proximate (< 132 in) ponds. EIowever, we were unable to determine whether juvenile oak toads immigrate to nonnatal ponds as adults fbr breeding. Juvenile recruitirxnt was Lippareintly low among all ponds, so "rescue" fro111 those ponds would h;ne heen rninii?~al if' it occurred. Instead, our recapture data suggest that oak toads some- times skipped breeding seasons at some ponds or during some years (e.g., 1997) and remained in the riplands surrounding the "extinct" ponds.

The 'ponds as patches" paradigm (Marsh and Tretlham, 2001), or metapopulation the- , supportstde mnseivation of multiple ponds within landscapes to permit local inter-pond extinctions and recolonizatioiis by amphibi;ms to occni-. We found little evidence that oak toads ftlnction as a metapopulation, although w e were unable to determine whether juveniles immigrate to nonnatal ponds for breeding as adults. We did, however, observe extreme variability in breeding effort and ju\ enile recruitment among ponds and years. Altlrc>ugh \vriitl~er- and pond hydi-0109 influenced pond use by <)ilk toads. those variihles fxiletl to explain rlnost of the variation i11 use imo1ig years ancl ponds. Clearly, kictors other tliarl tl~ose we tested inflrlence tlo~v ;arid wl~en oak toacis select ponds for breeding, and ~f hether ju-\ enilv recr~rittnent \w/ill be srtccess- fill. Kegartlless of' wlietlicr oak toads exhihit a ~twtapopnlatiolr structllre, their ci~nsew;itioi~ is hest ensnrecl b? mairltaining mnltiplc ponds ~;i.ithin a lantlscape. This nil1 enhance the probability of recrnitlnrnt n ithin the landscape neigl~l-tlorllool dlrsing ;it least sotnr years

Ac.Xnorclvcig?rtc~rtt'i -F~rntlr~ig \vas prolrtied by the CSD4 Forest Senlee, t~ielritltng the Ocala Ilatlonal Fore\t, the 1,otigleaf P ~ n e Ecos~stern Ke\toratrorr Pro- gr~irn, anci the Sotltlier1-t Kesearch Statloti's I3ent Creek Erpt.rilnrnta1 Forest, the llepartrrierit of Energ,- S,t-~,~nndi Nr\er Operation\ office thtol~glr tht' Forest Senlee S,nannall Rtver Site and the Forest Senice Soiithern Rr\earcli Statlo11 ~lrtdrr Iritet,tgerrci Agrrerrlerit I~E-4109-"iSK0OOt~6, and the Floritfa Frs11 ,~ncJ WrlcXhfe Conien~at~on Cotni~it\\lon, Bureaii of \.I rltllrfe llr\erstt]i Con\en,rtlotl, co11tr;tcts NC:99-014 arid C1195 We tl-rank J Beach for field and logtstrcal as\tstance, ,uld for her dedication to th14 \tud\ (: McMnhot~, D Loftrs, arid J Blake prov~cled crittc'il \rtpport to tlie ctndj S Mla/n>, D \?ionten, R Xslrtori, M Welker, J Smttlr, J Stager J L3asrchi\~ich, K Owen 1 ) Johnson, S Johii\on, J '\%iel)e, K Chrren, and other\ h'ive provrded ercellerit fielit asurtance ant1 otlrite ploject rurdr1,igeinerit We a150 thank the Ocala Nat~onal Fore\t \taff officers, rncli~ding L Lttliei-y, R Lowerv, C: SekeraL, J Clutts. J Marr M Clere, 'irid the Oealrt file crew for their a\sistance S Johnson and K K~rike~itl g,i\e vahral-rle snggest~oris for lrnplovlng an earlier verslorr of tlits maiiuscrtpt

AI FOIXI>, K A. 1989 ?'~i~tatlori trr precf'itor plie11010~ affect\ predator pe~formance and prry comrrrurirt> c o ~ t i p o ~ t ~ ~ n Ecolog~ 70.206-219

HI 1 ~ 1 5 7 ~ ~ h , A K.. 13 B II W.?rk~. ~ U I ) M' P Sob41 1994 Amph~bian cfeclines judg~ng ct,~.bilrt\i, per,tsterice, and st~rceptibilrty of populatrons to local ancl global extme- tion\ Concervatlorl Biology 8 60-7 1

(:\RN, A F , JH 1940 ?i contsik)utron to the herpetoloa iof Flortda Untverslt). of Florrda Publlc,it~ons (Biologicd Science) 3 1-113

I3ol>11 C K , Jli 1991 Drlft fence-assoc1,tted s,unplrng hias of amphrhrans <it ;I Florlda s'1ndlitl1 ternporarv poncl Jnrirrtal of ITerpetolop 25 2961-301

1992 Biologrcal diver\itv of d temporitry pond l~erpetofnnn~l In 11ort11 Flor~d~t sandhll Rtodiversitj and Conservatron 1 125-142

1994 Tlie effects of tlronght orr populatron \tructure, acttvtty and orientatrori of toads (B~i fo c/nci?-r-clc~i, and B tprri.rtmr) nt a tenrporan pontl Etl~olom E~eoloq and Evoliitton 6 33 1-,349

1996 U\e of terre\t~tal k~dbtt'lts. 1): d~iiphrh~dn\ 111

the sanclliill upl,lndr of tiortli-eer~tr'tl Florida ?Ilvtt>\ I4 42-52

(:HI- f. \ B E C: 11 1993 Effects of IIiglt-1rrte~isih \\ ll<lfire arid 5tlt1~11lttiral Treatrner~ts or1 Htotic (:o~rrrtlrtn~trc~\ of Sdrtti Ptne \crttb 1'11 13 Drssrrtation l l r l r t et s ~ t l of Florrdzr, C:arncs\ille, Florrda U 5 A

200 1 Sp."t")-tempor"1 d\-n,ttxirc s of pot1t1 tlse and t eel-trtttirent In F ~ O I - I ~ ~ I gopfier frt)gs iRi~ti(i ( ~ j 7 1 t c )

cl~rop?~d J~ottrn,il of IIerpetolo~ 35 71-h5 2002 Frt e llab~t~tt str uctur e a i d llerpetof,lnna tn

the sorrthe,i\t Pp 91-99 Irr ?? \1 Ford K K l-lrts\ell ,inti I: E \Ioorn~at~ I Ecls i Proceedings of the S\nrpow~in "Tlie Role o f File for Ilotigatne ?%'ilc1life t/l;llidgernent atltf C'or-tliir~rtiit\ Kestor,lt~c)n Tratltttoniil

User ji1d \em Dllectlon\ ' '7"\htinual Conference of the ?t'~ldltfe Socteb , 15 September 2000 zla\h\llle, Ten- nesFee USDh Fot r \ t Sen ice Geri Tech Rep 1 E-288

C ~ K E P \ B E EIC, C H , \ \ I ) R 'i%' 51210\5 I999 4ge, com- position, and \ t a d \trrictrtre of old-gro\\tli oak sttes In the Florida high pnle landscape ~rllpl~~atrons for eco- \\stern mariager~ient dnd re\toratlon h atlrral Ire&\ Journd 19 30-30

CKEE\BI.KC, C fI 121>G W TIZ\FII 2004 Bteedirlg pond selection and mo\ ement pattel ns bx Spadefoot Toacis (Scaphzopu\ holbrooktz) tn relnttoll to weather dntl edaph~c condlttons Jonrnal of Her-petolog ,?38 569-577

E l I ~ ~ ~ L T O L . \V J , JR 195.5 Votes on the ecoloa of the oal\ toad In Florld,~ Herpetologica 11 205-210

11 12x1, I . \VII D S I ~ I K F H I OFF 1997 Tlie inetapopnlatlon approach it5 hirtoq, conceptual dornai~i, and appllca- t~on to c o n ~ e n ~ ~ ~ t ~ o n Pp 5-26 I n 1 4 Hansla and tl E Gllpin (Ed5 ), kletapopulatlon B~ology 4cadeiiiic Press, San Diego, Cnlifomtd, U S A

LITT, A R , L I) PRO\ 1 uc 1113 R, G W TZNUI-H. \ \ I ) K FR 1 % ~ 2001 I-Ieq~etofaunal 1 esponsrc to restoration treatment5 of longleaf pine sandhi115 111 Floncl'~ Restoration Ecologv 9 462474

MAR~JI , D M , AVD P C T R ~ Y H - \ " ~ ~ 2001 Metcipopulation ciynarnic\ and amphibian con\en~~itlon C>ortsenation Biolohy 115 4 0 4 9

hlli 51r \LA VI' E , J R , ZNII J N L i ~ w 2002 Herpetofauna of a long-unburned \antihlll habltat in qoilth-central Florlda Flonda Sc~entt\t 65 35-50

Mot FR, P E , %YI> R F R ~ Z 1987 Wildlife values of small, isolated wetlands 111 tlie mutheastern coastal pla~n Pp 234-241 In R R Odum, K A Ricidleberger and J C Ozier iEds ), Proceednlgs of the 3"' 5 E Nongdrne ,md Endangered W~ldl~fe Svmposium G e o r p Department of Yatural Resources, Atlanta, Georga, U S A

Pr\r,~r, J G. 1998 Breeding biology of tlie gopher frog, Aana capzto, in we\tern Florlda Journal of Herpetoloa 32 217-223

PFCHMAUY, J H K , D E SCOTT Jli J W G113130h5, IN13

K D SL. M I I15C 1989 Influence of wetland I~vdro- period on d~verclty and abundance of metnrnorplio\ing

ju\ en11e cir~lpli~bians \\'etlancis Ecoloa dnd Manage- ment 1 3-1 1

PFTH i \ h i J \IT \ \ I > C h KKF \ZC 111 1999 Pond tadpoles wrth ge~ier~ilizetl morphologt is tt time to reconstder the~r ft~tictional roles rn ~tcjtlatic cornmunitles? Oecolog~a 120 621-631

Pf.TR%\k\ J \;i C: 4 K P \ \ ~ I > I , I\1>5 S ' V ~ C K R A ' ~ 2003 Response of amphib~atir to re\toration of d southern hppalachlan tvetlancl a long-term anal~sis of cosnmtlnit) dxnamics \Vetlands 23 1030-1042

S~ur,trst r-1, R D , D E SCOTT, J H K P F C H M ~ \ \ , AYI)

J 1%' GIBBO\~ 1996 Structure and dynamic\ of an ampli~bian cornmun~t) e~idence from a 16-year study of '1 natural pond Pp 217-248 Zn M L Cody and J A Smallwood (Eds ), Long-term Studles of Vertebrdte Co1iimui11tle\ Acadenlic Press, San Diego Cahforn~a, U S A

S ~ F I I I D K E E W'EKNLW, IND S A COKTCVRILHT 1999 Lorig-telm distr~but~orial dvnamlcs of a

<,

Michigan amphibian asseinblage. Ecology 80:2326- 2337.

SKEI~I,~', D K., K. L. YC'RE~~TCI,, E. E. WFRSER, A N D R. A. KFLYE I. 2003. Ectimating decline and distributional d~ange in amhhibians. Conservation Biology 17: 744-75 1.

SIYSCFI, U. 1992. Structure and dynamics of a natterjack toad ~iletapopulation (Bufo cnlrfnitci). Oecologia 90: 489499.

WILIJUR, H. M. 1987. Regulation of structure in complex systems: experimental temporary pond communities. Ecology 68: 1437-1452.

WRIGHT, A. H. 2002. Life-histories of tlie Frogs of the Okefinokee Swamp. Georgia. North American Salientia (Anura) No. 2. Cornell University Press, Ithaca, Kew York, U.S.A.

ZAR, J. EI. 1984. Biostatistical Analysis. Prentice Hall, Engle~vood Cliffs, New Jersey, U.S.A.

Acceytecl: 6 June 2005 ilsst>ciate Editor: Dean Adams

spatial and temporal ecology of oak toads (b l7f0 qi

Documents