LATITUDINAL VARIATION IN PALATABILITY OF SALT-MARSH PLANTS: WHICH TRAITS ARE RESPONSIBLE?

Transcription

1 Ecology, 83(12), 2002, by the Ecological Society of Ameica LATITUDINAL VARIATION IN PALATABILITY OF SALT-MARSH PLANTS: WHICH TRAITS ARE RESPONSIBLE? ERIN L. SISKA, 1,3 STEVEN C. PENNINGS, 1,4 TRACY L. BUCK, 1,5 AND M. DENNIS HANISAK 2 1 Univesity of Geogia Maine Institute, Saelo Island, Geogia USA 2 Maine Sciences Division, Habo Banch Oceagahic Institution, Fot Piece, Floida USA Abstact. Biogeogahic theoy edicts that intense consume ey inteactions at low latitudes should select fo inceased defenses of ey elative to high latitudes. In salt mashes on the Atlantic coast of the United States, a community-wide atten exists in which 10 secies of low-latitude lants ae less alatable to a divese suite of hebivoes than ae high-latitude consecifics. Examination of oximate lant taits (toughness, alatability of ola and la extacts, nitogen content) of high- and low-latitude consecifics of nine lant secies suggested that all these oximate taits had the otential to contibute to latitudinal diffeences in alatability of some lant secies. Southen lants wee toughe than then lants (five secies), had less alatable ola extacts (fou secies), and had lowe N content (six secies). Exeimental evidence linking taits to latitudinal diffeences in alatability was stongest fo ola extacts and lacking fo N content. Fo one lant secies, ne of the taits we studied coelated with latitudinal vaiation in alatability. Because alatability diffeences may change when moving fom fesh lants to feeze-died lants to lant taits, studies of latitudinal vaiation in feezedied lants o lant taits ae likely to unde- o oveestimate latitudinal vaiation in alatability of fesh lants. This study has begun to identify the oximate lant taits esonsible fo latitudinal vaiation in lant alatability in Atlantic coast salt mashes, but the ultimate evolutionay factos esonsible fo vaiation in these taits emain to be detemined. Key wods: Atlantic coast (USA) salt-mash lants; biogeogahic theoy; chemical defense; hebivoy; latitude and lant vaiation; alatability, affected by multile lant taits; alatability, lant vaiaton acoss latitude; lant chemical defenses, latitudinal vaiation; lant hebivoe inteactions. INTRODUCTION Biogeogahic theoy edicts that inceased consume essue at low latitudes should select fo inceased defenses of ey elative to high latitudes (MacAthu 1972, Bakus 1974, 1981, Bakus and Geen 1974, Geen 1977, Vemeij 1978, Jeanne 1979, Betness et al. 1981, Gaines and Lubchenco 1982, Louda 1982, Fawcett 1984, Heck and Wilson 1987, Coley and Aide 1991, Stiven and Gadne 1992, Jablonski 1993, Conin et al. 1997). Because of the logistical difficulties inheent in woking acoss geogahic distances, most studies of latitudinal vaiation in lant alatability have t diectly comaed the alatability of high- and lowlatitude lants (but see Steinbeg et al. 1991, Swihat et al. 1994, Pennings et al. 2001). Instead, most studies have examined vaiation in lant taits such as tough- Manuscit eceived 4 Octobe 2001; evised 23 Ail 2002; acceted 1 May Pesent addess: 596A Mt. Pleasant Road, Toonto, Ontaio M4S 2M8, Canada. ein siska@hotmail.com 4 Pesent addess: Deatment of Biology and Biochemisty, Univesity of Houston, Houston, Texas USA. 5 Pesent addess: Bauch Maine Field Laboatoy, P.O. Box 1630, Geogetown, South Caolina USA ness, nutitional content, o the divesity, identity and concentation of seconday metabolites (e.g., Levin 1976, Rodiguez 1977, Coley and Aide 1991), o have woked with ocessed (lyohilized and ulveized) lant mateial (e.g., Bolse and Hay 1996). These aoaches have ceated thee caveats in ou cuent undestanding of latitudinal vaiation in lant alatability. Fist, vaiation in a aticula tait (say, helic content) does t necessaily imly vaiation in alatability, because hebivoes may t be sensitive to vaiation in a aticula lant tait (Steinbeg and van Altena 1992, Steinbeg et al. 1995). Moeove, since many taits combine to oduce alatability, vaiation in one tait (say, helics) might be counteacted by oosite tends in athe (say, toughness) with the esult that oveall diffeence in alatability exists. Although this concen can be atially allayed by measuing as many lant taits as ossible (e.g., Coley and Aide 1991, Coley and Baone 1996), it is logistically difficult to measue evey single tait that might matte to hebivoes, and had to kw how to integate multile tait measues into a single index of alatability. Second, although the majoity of studies to date have been consistent with the hyothesis that lant defenses incease at low latitudes (Hay and Fenical 1988, Coley and Aide 1991, Bolse and Hay 1996), a significant

2 3370 ERIN L. SISKA ET AL. Ecology, Vol. 83, No. 12 faction of studies have found, o the oosite, tend (Van Alstyne and Paul 1990, Steinbeg 1992, Tagett et al. 1992, Byant et al. 1994, Swihat et al. 1994). Reconciling these divegent esults is difficult because few studies have integated diect measues of lant alatability with studies of lant taits that might affect alatability. Thid, many geogahic studies have comaed unelated gous of lants, aising the ossibility that hylogeny, athe than latitude, is the imay dive of alatability (o taits). This concen can be avoided by comaing congenes within multile genea (Bolse and Hay 1996) o making intasecific comaisons (Bolse and Hay 1996, Pennings et al. 2001). Salt mashes offe an attactive system fo examining latitudinal vaiation in lant alatability using aoaches that avoid the above caveats. Salt mashes ae a dominant intetidal habitat along the Atlantic coast of the United States (Pennings and Betness 2001). The lant community of coastal salt mashes is comised of elatively few secies (Chaman 1974), making it logistically ossible to study vitually the entie community. In addition, most of the lant secies occu acoss a wide latitudinal ange (Duncan and Duncan 1987, Pennings and Betness 1999), making it ossible to make intasecific comaisons acoss latitude. We studied latitudinal vaiation in lant alatability in salt mashes on the Atlantic coast of the United States, woking with the majoity of the lant and hebivoe secies esent in the community, comaing lants fom New England with consecifics fom the South Atlantic Bight. We combined diect measues of lant alatability with measuements and exeimental studies of lant taits in ode to detemine which taits might be esonsible fo obseved vaiation in alatability. Results of alatability tials with fesh lants wee ublished eviously (Pennings et al. 2001), and demonstated stong efeences fo then vs. southen individuals of all 10 lant secies studied, egadless of season of lant collection o geogahic oigin of hebivoe used. Hee, we assess the elative imotance of lant toughness, chemical defenses, and N content in oducing this stiking geogahic atten. Ou exeiments aallel those used by Bolse and Hay (1996) to study latitudinal vaiation in the alatability of seaweeds, and we daw comaisons between the esults of the two studies in the discussion. METHODS Study sites To ensue that woking at any one site did t bias the esults, we collected lants fom fou then (Rhode Island, USA) and six southen (Geogia, USA) sites. The then sites, located nea N latitude, wee Hunded Ace, Rumstick, Haffeneffe, and Long Neck Cove, coded H, R, F, and N, esectively, in the figues. The southen sites, located nea N latitude, wee Aiot, Shell, Island, Lighthouse, Mash Landing, and Cabetta, coded A, S, I, L, M, and, C, esectively, in the figues. Long Neck is a otected cove of the Sakonnet Rive, nea Common Fence Point, Rhode Island. All othe sites ae descibed in Pennings et al. (2001). Plants and hebivoes We woked with most lant secies common at both then and southen sites: Aste tenuifolius; Iva futescens and Solidago semeviens (Asteaceae); Distichlis sicata, Satina altenifloa, and Satina atens (Poaceae); Saliconia euoaea and Saliconia viginica (Cheodiaceae); and Limonium caolinianum (Plumbaginaceae). These nine secies eesented the majoity of the secies and 75% of the total lant biomass in both geogahic locations (authos esonal obsevation), and included 9 of the 10 secies studied by Pennings et al. (2001). Feeding tials wee conducted using thee gasshoe and one cab secies, selected because they wee common at Saelo Island, Geogia, and adated well to laboatoy conditions. The majoity of the tials wee conducted with the cab Amases cineeum (Gasidae), coded Ac in the figues, because it was highly omnivoous and thus facilitated comaisons by allowing us to examine all the lants using a single consume. A small numbe of additional tials wee conducted with the gasshoes Ochelimum fidicinum (Tettigoniidae), Ohulella elidna (Acididae), and Paoxya clavulige (Acididae), coded Of, O and Pc esectively in the figues, to confim that esults fom cab tials wee tyical of othe consumes. In evious wok, both then and southen consumes efeed then vs. southen lants, and esults fom cab tials wee simila to esults fom 12 othe consumes (Pennings et al. 2001). See Pennings et al. (2001) fo additional infomation on the lants and consumes. To ensue that ou esults did t vay ove the gowing season, we collected lants twice in each of two yeas (June and ealy Setembe of 1997 and of 1998) fo toughness, nutient, and helic analyses. Feeding tials wee conducted using lants collected in We will efe to June and Setembe as ealy and late seasons, esectively. Toughness measuements To comae the toughness of then and southen lants, we measued toughness using two diffeent techniques. These techniques, the enetomete test and the iing test, mimic diffeent modes of hebivoy, give simila ankings among lant secies, and ae descibed in Pennings et al. (1998). Biefly, the enetomete test measues the foce needed to enetate a leaf with a thin od, and the iing test measues the foce needed to i a in though a vetically oiented leaf. We did t use the iing test with D. sicata

3 Decembe 2002 SALT-MARSH PLANTS: LATITUDINAL VARIATION 3371 and S. atens because the foce equied to i these lants was below the detection limit of the aaatus. Fo each test, individual leaves of each lant secies wee collected fom each of two sites, etuned on ice to the laboatoy, and measued within 4 h. Values fo individual leaves wee aveaged to yield a single value e secies e site e season. Latitudinal diffeences in toughness wee examined using a theeway ANOVA, with sites as elicates and latitude, season and yea as main effects. Data fo Saliconia viginica wee analyzed with a two-way ANOVA because lants of this secies wee collected ealy in Data wee log tansfomed when aoiate to satisfy assumtions of ANOVA. Reconstituted-lant feeding tials To detemine if vaiation in toughness influenced vaiation in alatability, we offeed consumes a choice between econstituted-lants fom each geogahic egion. We easoned that beaking down the hysical stuctue of lants and econstituting them into atificial diets would allow a comaison of lant chemical comosition in the absence of diffeences in toughness. Thus, if consumes longe disciminated between then and southen lants in econstituted fom, it would suggest that toughness layed a ole in diffeences in alatability of fesh lants. This conclusion, howeve, would come with the caveat that chemical changes might have occued duing the ocessing of the lant mateial (see Discussion, below). Altenatively, if consumes still efeed econstituted then lants ove southen consecifics, it would suggest that diffeences in lant chemisty wee esonsible fo diffeences in alatability. Plants wee fozen, lyohilized, and gound to a owde using a Wiley mill (mesh size 40 m). Plant owde fom a then site was econstituted (0.2 g/ ml) into a 4% aga diet and oued into one of two ectangula temlates esting on mesh sceen (mesh size 1mm 2 ). Consecific lant owde fom a southen site was similaly econstituted and oued into the second ectangula temlate. Afte the aga solidified, stis of mesh sceen containing aied diets wee esented to consumes in a two-choice feeding tial. Consumes wee housed within elicate glass jas containing wate and the two consecific diets. Relicates (n 20) wee checked twice daily and teminated when substantial consumtion of at least one diet occued. Relicates in which neithe diet was eaten afte 72 h, o in which both diets wee comletely consumed between obsevations, ovided infomation on the elative alatability of then vs. southen diets, and wee omitted fom the analysis. Consumtion was measued as the aea of each diet consumed (in squae millimetes). Diffeences in consumtion wee comaed between geogahic locations using aied t tests, o, when assumtions of mality o equal vaiance wee t met, with Wilcoxon signedank tests. Chemical extact feeding tials To diectly detemine if diffeences in lant chemisty contibuted to latitudinal vaiation in alatability, we comaed the alatability of then and southen lant extacts. Two solvent mixtues wee used in the extactions in ode to obtain comounds with a wide ange of olaity. Plants wee collected and fozen. Aoximately 1 kg of lant mateial fom each site in each season was maceated and extacted twice in 2:1 dichloomethane : methal, followed by thee extactions in 70:30 methal : wate. Each extaction used 300 ml of combined solvents. The solvent mixtues fom the five extactions wee combined and atitioned in a seaatoy funnel into one ola (methal : wate) and one la (dichloomethane) extact. The extacts wee educed to dyness using otay evaoation and lyohilization. Pola and la extacts wee tested seaately. Extacts fom then and southen lants wee incooated at natual concentations (mass e volume) into a 4% aga diet, and esented to hebivoes in twochoice feeding tials; tials wee conducted and data analyzed as descibed above (see Reconstituted-lant feeding tials). Because we lacked detailed infomation on how vaious asects of nutitional quality vaied among lant secies and geogahic egions, we chose to incooate the extacts into an aga diet that lacked additional nutitional constituents. The aga extact diets wee eadily eaten by the consumes; howeve, esults should be inteeted with caution because consumes may be moe sensitive to seconday metabolites in diets with low nutitional quality (Duffy and Paul 1992, Hay et al. 1994). Nitogen and helic concentations Because geogahic diffeences in lant N content might mediate alatability, we measued total N content of then and southen lants. Plants wee collected, lyohilized, and ulveized using an amalgamato. CHN analysis was efomed at the Univesity of Geogia Chemical Analysis laboatoy using a Pekin- Elme 240C elemental analyze (Pekin-Elme, Wellesley, Massachusetts, USA). Nitogen content (ootion of dy mass) was acsine (squae-oot) tansfomed and comaed fo each secies with a thee-way AN- OVA (latitude season yea). Individual lants (n 4) wee nested within sites (n 2 lants e egion e season e yea), which wee the unit of elication. Fo most of the lant secies that we studied, bioassay-guided factionation exeiments have t been efomed to identify the seconday metabolite(s) that mediate alatability. In the case of Satina altenifloa, howeve, thee is evidence that the imay seconday metabolites ae simle helics such as feulic and coumeic acid (Valiela et al. 1979, Buchsbaum et

4 3372 ERIN L. SISKA ET AL. Ecology, Vol. 83, No. 12 al. 1984, Rietsma et al. 1988, Newell and Pote 2000). We measued total helic content of S. altenifloa leaves using a standad Folin-Dennis eaction (Julkunen-Tiitto 1985, Zimme 1997). Leaves wee fozen afte collection, lyohilized, and ulveized. Fifty milligams (dy mass) of leaf mateial was extacted in 80% methyl alcohol 3 h, agitating extacts evey 30 min. Fifty micolites of extact suenatant wee eacted with 750 L deionized H 2 O, 200 L Folin-Ciocalteu eagent, and 1000 L 20% Na 2 CO 3. The absobance at 700 nm was measued on a sectohotomete 20 minutes afte addition of Na 2 CO 3. Concentations wee calculated based on a standad cuve using ue feulic acid (Sigma Chemical Comany, Saint Louis, Missoui, USA). Because Folin-Dennis measuements have a wide vaiety of limitations (Ael et al. 2001), esults should be inteeted with caution. RESULTS Toughness measuements Five lant secies wee toughe in the south than in the th, based on at least one of the two toughness measuements (Fig. 1). In case did the data suggest that then lants wee consistently toughe than southen consecifics. Penetomete test. Most of the secies tested, with the excetion of Aste tenuifolius and Solidago semeviens, vaied in this measue of toughness as a function of latitude, season, o yea; howeve, inteactions between main effects often comlicated unambiguous identification of latitudinal attens (Table 1). Nevetheless, insection of the data (Fig. 1) evealed that southen Satina altenifloa, S. atens, Limonium caolinianum, and Saliconia viginica wee toughe than then consecifics on moe than half of the dates. Latitudinal vaiation in toughness of Satina atens was aticulaly stiking, vaying as much as thee-fold between southen and then lants. Riing test. Thee of the lants tested, A. tenuifolius, S. altenifloa and Saliconia viginica, vaied in this measue of toughness as a function of latitude. In the case of A. tenuifolius the magnitude of this diffeence vaied between yeas (Table 1); howeve, southen lants had highe toughness measuements than then lants on each date (Fig. 1). Latitudinal vaiation in toughness of Satina altenifloa was aticulaly stiking, with then lants iing at half the foce of southen lants. Reconstituted-lant feeding tials Twenty-six out of 47 (55%) tials with econstituted lants esulted in geate consumtion of the then diet than of the southen diet (Fig. 2). In only one econstituted tial was a southen lant efeed ove a then consecific. The natue of the esults vaied between lant secies, with five secies dislaying stong latitudinal tends, two dislaying inconclusive esults, and two dislaying latitudinal tends. Gasshoes efeed the then diet in 8 of 11 tials and cabs in 18 of 36 tials (P 0.30, Fishe exact test). Ove half of the tials with Iva futescens (5 of 7), Satina altenifloa (5 of 7), S. atens (5 of 6), Saliconia euoaea (3 of 4) and S. viginica (3 of 4) esulted in significantly geate consumtion of the then lant than the southen consecific. Two additional tials wee maginally significant (P 0.06), and all the tials fo these five secies showed a tend towads geate consumtion of the then diet. Consequently, these assays ovided stong evidence that, fo these five secies, thee ae latitudinal diffeences in lant quality othe than, o in addition to, diffeences in toughness. Half of the tials with Aste tenuifolius (2 of 4) and Limonium caolinianum (2 of 4) esulted in significantly geate consumtion of the then lant than the southen consecific. In the case of A. tenuifolius, esults deended uon which then and southen sites wee aied (e.g., one ai of sites dislayed significant diffeences on both dates; the othe ai did t). In the case of L. caolinianum, the esults deended on season, with both ealy tials nsignificant, and both late tials significant. Fo both secies, at least one tial showed a nsignificant tend towads geate consumtion of southen lants. Because of the vaiability in these esults, we conside them suggestive but t conclusive evidence fo latitudinal diffeences in lant quality unelated to toughness. None of the tials with Distichlis sicata (0 of 6) and only one of the tials with Solidago semeviens (1 of 5) esulted in significantly geate consumtion of the then lant than the southen consecific. Moeove, one of the tials with S. semeviens esulted in significantly geate consumtion of the southen lant. Fo these secies, the econstituted-lant tials ovided evidence of latitudinal diffeences in lant quality unelated to toughness. Chemical extact feeding tials Oveall, 6 out of 37 (16%) tials with la extacts and 23 of 37 (62%) tials with ola extacts esulted in significantly geate consumtion of the then than the southen diet (Fig. 3). Only one tial with a la extact and one tial with a ola extact esulted in significant efeences fo the southen diets. Results vaied among lant secies, with fou showing conclusive evidence fo latitudinal diffeences in alatability of the ola extact, thee showing suggestive but inconclusive esults, and two dislaying latitudinal tends. Ove half the tials with ola extacts of Iva futescens (5 of 5), Satina altenifloa (4 of 4), S. atens (3 of 4, counting one tial with P 0.054), and Saliconia euoaea (3 of 4) esulted in geate consumtion of the then than the southen diet. All nsignificant tials with ola extacts of these secies

5 Decembe 2002 SALT-MARSH PLANTS: LATITUDINAL VARIATION 3373 FIG. 1. Toughness of then and southen lants, tested by enetomete and iing methods. Data ae means and 1 SE. Season and yea of collection ae indicated below each ba.

6 3374 ERIN L. SISKA ET AL. Ecology, Vol. 83, No. 12 TABLE 1. Plant toughness: thee-way ANOVA esults of enetomete () and iing () tests. Aste tenuifolius Iva futescens Secies Test Latitude Season Yea Solidago semeviens Distichlis sicata Satina altenifloa S. atens Saliconia euoaea S. viginica Limonium caolinianum Latitude Season Note: Significant esults (P 0.05) ae indicated in boldface. Measued only with enetomete. Evaluated with two-way ANOVA because measuements wee made in ealy Latitude Yea Season Yea Latitude Season Yea showed a tend towads geate consumtion of the then diet. In addition, two of five tials with la extacts of I. futescens also esulted in geate consumtion of the then than the southen diet. All fou of these lant secies eviously dislayed stong latitudinal attens in alatability of econstituted lants, and we conclude that this ealie esult was imaily diven by latitudinal vaiation in alatability of ola extacts. The econstituted-lant assays ovided stong evidence fo latitudinal diffeences in alatability of Saliconia viginica, and inconclusive evidence fo Aste tenuifolius and Limonium caolinianum. In the chemical-extact tials with these secies, half of the tials with ola extacts esulted in geate consumtion of the then than the southen diet (two out of fou in each case). In the case of A. tenuifolius, the esults again diffeed as a function of which sites wee comaed, but the ai of sites that consistently diffeed in alatability in the econstituted-lant assays (Rumstick and Aiot) did t diffe in the ola-extact assays, and the ai of sites that consistently diffeed in alatability in the ola-extact assays (Hunded Ace and Mash Landing) did t diffe in the econstituted assays. Similaly, esults fo L. caolinianum again diffeed by season, but wheeas the econstituted-lant assays dislayed latitudinal diffeences only in the late season, the ola-extact assays dislayed diffeences only in the ealy season. We conside these esults suggestive but t conclusive evidence fo latitudinal vaiation in lant alatability diven by extacts. Thee was evidence in the econstituted-lant tials fo latitudinal diffeence in alatability of Solidago semeviens o Distichlis sicata, did these secies dislay stong evidence fo latitudinal vaiation in extact alatability. Two of five assays with ola extacts of S. semeviens esulted in geate consumtion of the then than the southen diet. Results with D. sicata wee mixed, with one assay indicating a efeence fo the then diet, one a efeence fo the southen diet, and two efeence. Nitogen and helic concentations Nothen oulations of seveal lant secies had highe N concentations than did southen consecifics (Fig. 4), but the magnitude of the diffeences often vaied as a function of season o yea, leading to significant inteactions between the main effects (Table 2). Insection of the data indicated that Aste tenuifolius, Distichlis sicata, Satina altenifloa, Saliconia euoaea, S. viginica and Limonium caolinianum had highe levels of N in the th than in the south on at least thee of fou dates. Diffeences wee often lage ealy in the yea than late in the yea, leading to significant latitude season inteactions (Table 2). At least thee additional secies, Iva futescens, Solidago semeviens, and Satina atens, had highe levels of nitogen in the th than in the south ealy in the yea, but showed minimal diffeences o the evese atten late in the yea. Given the shote gowing season in the th, these esults ae consistent with young, aidly gowing lants having high N content, and olde, senescing lants having low folia N content because of tanslocation to ovewinteing ogans. In any case, because these secies had a highe N content than southen consecifics only in the sing, latitudinal vaiation in N content might exlain diffeences in alatability that wee obseved in the sing, but could t exlain diffeences in alatability that wee obseved in the fall (Pennings et al. 2001: Fig. 2).

7 Decembe 2002 SALT-MARSH PLANTS: LATITUDINAL VARIATION 3375 FIG. 2. Consumtion of econstituted then vs. southen lants in two-choice feeding tials. Data ae means and 1 SE. Samle sizes, sites, and hebivoe used ae indicated below aied bas. Fo site abbeviations see Methods: Study sites. Hebivoe key: Ac Amases cineeum, Of Ochelimum fidicinum, O Ohulella elidna, Pc Paoxya clavulige. Phelic concentations of Satina altenifloa wee almost 30% geate in the south ( % of dy mass [mean 1 SE]) than in the th ( %, P , ANOVA, esults did t vay by yea o season). This esult is consistent with the latitudinal diffeences in alatability of ola extacts obseved in most comaisons with this secies (Fig. 3). Oveall, we found stong evidence that all secies of salt mash lants demonstate latitudinal vaiation in alatability when esented to hebivoes in an intact fom (Pennings et al. 2001). When we examined individual lant taits, we found that all but one secies exhibited latitudinal vaiation in at least one tait (Table 3). Five secies wee toughe in the south than the th. Five secies wee less alatable in the south than the th when tested in econstituted fom. The ola extacts of fou of these five secies wee less alatable in the south than the th. Finally, nitogen content was lowe fo six secies, and helic concentation highe fo one secies, in the south than the th. DISCUSSION Along the Atlantic coast of the United States, southen salt-mash lants ae less alatable than then consecifics (Pennings et al. 2001). Hee, we have begun to identify the lant taits that ae esonsible fo this vaiation. Plant toughness, seconday chemisty, and nitogen content all vaied with latitude in some lant secies, but exeimental evidence linking tait vaiation to vaiation in alatability was stongest fo

8 3376 ERIN L. SISKA ET AL. Ecology, Vol. 83, No. 12 FIG. 3. Results of feeding tials with ola and la lant extacts. Data (consumtion of then vs. southen diets in two-choice feeding tials) ae means and 1 SE. Samle sizes, sites, and hebivoe used ae indicated below aied bas. Fo site abbeviations see Methods: Study sites. Hebivoe key: Ac Amases cineeum, Pc Paoxya clavulige.

9 Decembe 2002 SALT-MARSH PLANTS: LATITUDINAL VARIATION 3377 FIG. 3. Continued. ola extacts. Results with econstituted lants and lant taits wee less stiking than evious esults with fesh lant mateial, suggesting that studies that wok with econstituted lants, o that focus only on lant taits, ae likely to undeestimate vaiation in lant alatability. Many of ou esults ae simila to esults fom a latitudinal comaison of seaweed alatability by Bolse and Hay (1996), suggesting that simila ocesses may occu in both systems. In at least five salt-mash lant secies, southen individuals wee toughe than then consecifics in at least one measue of toughness. Fo ne of these secies did latitudinal diffeences in alatability totally disaea when lants wee comaed in econstituted fom; howeve, in two cases (Aste tenuifolius and Limonium caolinianum) the econstituted assays wee ambiguous, suggesting that emoving diffeences in toughness emoved some of the diffeences in alatability. Nevetheless, the thee lants with the stongest diffeences in toughness (Saliconia viginica, Satina altenifloa, and S. atens) also showed clea diffeences in alatability in econstituted assays when diffeences in toughness wee emoved. Thus, latitudinal vaiation in toughness was t the only tait leading to latitudinal vaiation in alatability; howeve, latitudinal vaiation in toughness could have einfoced alatability diffeences caused by othe lant tait(s). Pennings et al. (1998) found that vaiation in toughness was the imay facto affecting feeding efeences of Amases cineeum, the cab used hee in most of ou assays; howeve, thei study focused on comaing diffeent lant secies which vaied in toughness by 2 odes of magnitude. In comaison, the intasecific diffeences in toughness acoss latitude that we eot hee wee modest (u to 3-fold). We conclude that although intasecific vaiation in toughness could have contibuted to latitudinal diffeences in alatability, it was t of oveiding imotance. In at least fou and as many as seven of the lants (deending on the igo of the evidence demanded), latitudinal vaiation in the alatability of ola extacts contibuted to latitudinal vaiation in alatability of fesh lants. Fo all but two of these seven secies, which wee ambiguous fo both tests, latitudinal diffeences in alatability wee also found in the econstituted-lant assays. In the one case in which we quantified vaiation in a class of seconday metabolites, the helic concentation of Satina altenifloa was highe in southen lants than in then consecifics. Thus, fo these secies we can link vaiation in alatability of fesh lants to vaiation in alatability of econstituted lants (indicating that factos in addition to toughness ae imotant) to vaiation in alatability of ola extacts and in one case, otentially to vaiation in the concentation of helics. Documenting latitudinal vaiability in concentations of individual seconday metabolites must await bioassay-guided identification of which seconday metabolites in each lant ae distasteful to consumes. Moeove, because helic assays measue a boad gou of comounds that may diffe in activity and include comounds that ae t tue helics (Ael et al. 2001, Kubanek et

10 3378 ERIN L. SISKA ET AL. Ecology, Vol. 83, No. 12 FIG. 4. Nitogen content of then and southen lants. Yea and season of collection ae esented below each ba. al. 2001), the helic measuements should be consideed eliminay until confimed by quantification of individual comounds followed by feeding assays with comounds at natual concentations. Bolse and Hay (1996) similaly found that the alatability of extacts coelated with alatability of econstituted seaweeds bette than did othe seaweed taits. Nitogen content in seveal lant secies was highe in the th than the south. This vaiation could aise fom seveal souces. Notheasten estuaies ae tyi- TABLE 2. Plant nitogen content: esults of thee-way ANOVAs. Secies Latitude Season Yea Latitude Season Latitude Yea Season Yea Latitude Season Yea Aste tenuifolius Iva futescens Solidago semeviens Distichlis sicata Satina altenifloa Satina atens Saliconia euoaea Saliconia viginica Limonium caolinianum Note: Significant esults (P 0.05) ae indicated in boldface.

11 Decembe 2002 SALT-MARSH PLANTS: LATITUDINAL VARIATION 3379 Summay of esults of tests fo lant taits esonsible fo latitudinal vaiation in alatibity of nine salt-mash lant secies: indicates that we unambiguously identified latitudinal vaiation in a tait, maybe indicates esults wee equivocal, and indicates that we found little evidence fo latitudinal vaiation. TABLE 3. Satina Solidago Saliconia Limonium Aste Iva semeviens Distichlis alte- caolini- tenuifolius futescens sicata nifloa atens euoaea viginica anum Palatability of fesh lants Toughness Penetomete Riing maybe n/a n/a maybe Palatability of econstituted lants maybe maybe Palatability of extacts Nola Pola maybe maybe maybe Nitogen content maybe maybe maybe Notes: An n/a enty indicates that the tait was t measued fo this secies. Results fo fesh lants ae fom Pennings et al. (2001). cally moe eutohic than southeasten ones (Bicke et al. 1999). The shote gowing season in then mashes might equie a highe N content to allow fo aid gowth. Finally, low N content of southen lants might be an adative esonse to educe alatability to hebivoes (McNeill and Southwood 1978, Den and McClue 1983, White 1984, 1993, Augne 1995, but see Valentine and Heck 2001). Whethe latitudinal vaiation in N contibutes to vaiation in lant alatability is unclea, fo thee easons. Fist, Amases cineeum, the consume used in most of ou studies, does t aea to esond stongly to vaiation in lant N content (Pennings et al. 1998). Second, diffeences in N content wee most ounced ealy in the gowing season, but diffeences in alatability wee obseved both ealy and late (Pennings et al. 2001). Thid, we did t efom exeiments in which we maniulated N content in ode to assess consume esonses. Thus, ou data ae consistent with the hyothesis that diffeences in N content contibute to latitudinal vaiation in alatability of fesh lants, but do t diectly test this hyothesis. In comaison, Bolse and Hay (1996) uled out otein content as an exlanation fo latitudinal diffeences in alatability of seaweeds. A otential caveat to ou esults is that most of ou assays wee conducted with a single consume secies that is found only in southen mashes. It is ossible that athe secies of consume might have eacted diffeently in ou assays. Fo examle, a then consume might have been moe sensitive to vaiation in seconday chemisty (Conin et al. 1997). In ou evious study of fesh lants (Pennings et al. 2001), howeve, a suite of 13 consumes, including Amases, exhibited simila attens of efeence fo then vs. southen lants, egadless of consume secies o geogahic oigin. Moeove, esults in ou esent ae of 14 assays with thee secies of southen gasshoe wee qualitatively simila to esults of aallel Amases assays. Thus, although the use of diffeent consumes might have alteed some of the details of ou esults, the oveall attens and conclusions likely would t have changed. Oveall, as we moved fom fesh lants to econstituted lants to extacts, latitudinal diffeences in alatability deceased. Thus, all nine secies of lants diffeed in alatability in assays with fesh lant mateial, but only five unambiguously diffeed in alatability in econstituted assays, and only fou in assays with extacts. Fo one secies, Solidago semeviens, taits diffeed acoss latitude. Similaly, Bolse and Hay (1996) wee only able to tace latitudinal diffeences in alatability of econstituted seaweeds to extacts in about two thids of the cases. Thee exlanations might account fo these damening diffeences. Fist, if moe than one lant tait contibutes to diffeences in alatability, tests that eliminate some lant taits will find less stiking diffeences. Fo examle, if diffeences in alatability ae atially diven by factos othe than seconday metabolites, this at of the vaiation will t be esent in tests with extacts. Second, each time lant mateial is ocessed, vaious chemical comonents may volatilize, decomose, o eact with each othe. Thus, each ste in handling the lant mateial may destoy o dilute the taits of inteest. Finally, vaiation in alatability could be atially caused by taits that wee t investigated hee. Many taits combine to detemine lant alatability, and vaiation in taits such as the content of wate, salt, silica, o nutients othe than nitogen could have contibuted to latitudinal gadients in alatability. Regadless of the exlanation, this atten suggests that studies that wok with econstituted lants, o look only at lant taits, may undeestimate latitudinal diffeences in alatability of fesh lants. As discussed ealie (see Intoduction) it is also ossible that studies that focus on lant taits may oveestimate latitudinal diffeences in alatability if the studied taits do t affect alatability.

12 3380 ERIN L. SISKA ET AL. Ecology, Vol. 83, No. 12 One lant tait that we did t study that might vay with latitude is salt content. Southen mashes ae tyically saltie than then mashes because of inceased evaotansiation at lowe latitudes (Pennings and Betness 1999, 2001). To the extent that the salt concentation of lant tissues eflects that of the soil, southen lants might be saltie than then ones (howeve, in ilot studies, latitudinal diffeences in ash content wee neithe lage consistent). High salt concentation might dete feeding by some hebivoes, but t by Amases, which efes a high-salt diet (Pennings et al. 1998). Moeove, feeding efeences by Amases fo fesh lant mateial did t diffe fom those of othe consumes (Pennings et al. 2001). Thus, although we lack exeimental data evaluating the ole that salt concentation might lay in feeding efeences of most mash hebivoes, we doubt that it is a geneal exlanation fo the latitudinal attens in alatability obseved by Pennings et al. (2001). No single lant tait vaied in a way that was consistent with latitudinal vaiation in lant alatability fo evey secies. Rathe, all the taits vaied with latitude in seveal, but t all, lants. This divesity of tait diffeences acoss latitude is t suising. Because alatability is detemined by many taits, lants can be unalatable in diffeent ways. Because a wide vaiety of consumes including sa-sucking insects, leaf-chewing insects, cabs, snails, and mammals eat salt mash lants (Pennings and Betness 2001), diffeent t of defenses may be selected fo by diffeent consumes. Moeove, abiotic factos such as edahic o climate conditions also vay with latitude and may act concuently with consumes to exet selective essue on lants. Plant esonses to edahic conditions o climate may indiectly affect taits imotant to consumes, futhe inceasing the divesity of mechanisms esonsible fo latitudinal gadients in lant alatability. In this ae we have evaluated the otential fo lant toughness, seconday chemisty, and N content to detemine latitudinal attens of lant alatability. Although a geat deal moe wok emains to be done (we lack exeimental studies addessing the ole of N content, and we lack kwledge about the seconday metabolites esent in each lant secies), ou esults ae a fist ste towads exlaining the oximate lant taits that contibute to latitudinal vaiation in lant alatability in Atlantic coast salt mashes. The ultimate evolutionay mechanisms behind this latitudinal vaiation in lant alatability and lant taits, howeve, emain to be investigated. Ou wok has been esented in the context of standad biogeogahic theoies aguing that consume ey inteactions ae moe intense at lowe latitudes, and hence that ey defenses should be bette develoed at lowe latitudes (Mithen et al. 1995). Some cicumstantial evidence suots the hyothesis that hebivoe essue is geate in low- than high-latitude salt mashes (Pennings et al. 2001), but a vaiety of othe factos, including climate, length of gowing season, salinity, and anthoogenic eutohication also may diffe between low- and high-latitude salt mashes, and could lay a ole in latitudinal diffeences in lant taits and alatability. Futue wok by ou gou will evaluate these ossibilities in moe detail. ACKNOWLEDGMENTS We thank Rebecca Wisdom, Joan Maie Svedbeg, Heathe Gove, May-Besto Gant, Auda Lusche, Chistine Mille, Kisten Stohm, Dain Mooe, and Kut Reinhat fo hel with consume collection, lant collection, and setu of feeding tials. This wok was suoted by a Sigma Xi Gant-In- Aid (to E. L. Siska), the Univesity of Geogia Maine Institute Inten Pogam (E. L. Siska), and The National Geogahic Society (S. C. Pennings). A otion of this wok was included in a thesis submitted to the Deatment of Biological Sciences, Floida Institute of Techlogy, in atial fulfillment of the equiements fo the M.S. degee of E. L. Siska (M. Dennis Hanisak, committee chai). This is contibution numbe 903 fom the Univesity of Geogia Maine Institute and contibution numbe 1476 fom Habo Banch Oceagahic Institution. LITERATURE CITED Ael, H. M., H. L. Gove, M. D Ascenzo, E. Siska, and J. C. Schultz Limitations of folin assays of folia helics in ecological studies. Jounal of Chemical Ecology 27: Augne, M Low nutitive quality as a lant defence: effects of hebivoe-mediated inteactions. Evolutionay Ecology 9:1 12. Bakus, G. J Toxicity in holothuians: a geogahical atten. Biotoica 6: Bakus, G. J Chemical defense mechanism on the Geat Baie Reef, Austalia. Science 211: Bakus, G. J., and G. Geen Toxicity in songes and holothuians: a geogahical esective. Science 185: Betness, M. D., S. D. Gaity, and S. C. Levings Pedation essue and gastood foaging: a toical-temeate comaison. Evolution 35: Bolse, R. C., and M. E. Hay Ae toical lants bette defended? Palatability and defenses of temeate vs. toical seaweeds. Ecology 77: Bicke, S. B., C. G. Clement, D. E. Pihalla, S. P. Olando, and D. R. G. Faow National estuaine eutohication assessment: effects of nutient enichment in the nation s estuaies. NOAA, National Ocean Sevice, Secial Pojects Office and the National Centes fo Coastal Ocean Science, Silve Sing, Mayland, USA. Byant, J. P., R. K. Swihat, P. B. Reichadt, and L. Newton Biogeogahy of woody lant chemical defense against swshoe hae bowsing: comaison of Alaska and easten Noth Ameica. Oikos 70: Buchsbaum, R., I. Valiela, and T. Swain The ole of helic comounds and othe lant constituents in feeding by Canada geese in a coastal mash. Oecologia 63: Chaman, V. J Salt mashes and salt desets of the wold. Intescience Pess, New Yok, New Yok, USA. Coley, P. D., and T. M. Aide Comaison of hebivoy and lant defenses in temeate and toical boad-leaved foests. Pages in P. W. Pice, edito. Plant animal inteactions: evolutionay ecology in toical and temeate egions. John Wiley & Sons, New Yok, New Yok, USA. Coley, P. D., and J. A. Baone Hebivoy and lant defenses in toical foests. Annual Review of Ecology and Systematics 27:

13 Decembe 2002 SALT-MARSH PLANTS: LATITUDINAL VARIATION 3381 Conin, G., V. J. Paul, M. E. Hay, and W. Fenical Ae toical hebivoes moe esistant than temeate hebivoes to seaweed chemical defenses? Diteeid metabolites fom Dictyota acutiloba as feeding deteents fo toical vesus temeate fishes and uchins. Jounal of Chemical Ecology 23: Den, R. F., and M. S. McClue Vaiable lants and hebivoes in natual and managed systems. Academic Pess, New Yok, New Yok, USA. Duffy, J. E., and V. J. Paul Pey nutitional quality and the effectiveness of chemical defenses against toical eef fishes. Oecologia 90: Duncan, W. H., and M. B. Duncan Seaside lants of the Gulf and Atlantic coasts. Smithsonian Institution Pess, Washington, D.C., USA. Fawcett, M. H Local and latitudinal vaiation in edation on an hebivoous maine snail. Ecology 65: Gaines, S. D., and J. Lubchenco A unified aoach to maine lant hebivoe inteactions. II. Biogeogahy. Annual Review of Ecology and Systematics 13: Geen, G Ecology of toxicity in maine songes. Maine Biology 40: Hay, M. E., and W. Fenical Maine lant hebivoe inteactions: the ecology of chemical defense. Annual Review Ecology and Systematics 19: Hay, M. E., Q. E. Kael, and W. Fenical Synegisms in lant defenses against hebivoes: inteactions of chemisty, calcification, and lant quality. Ecology 75: Heck, K. L., J., and K. A. Wilson Pedation ates on decaod custaceans in latitudinally seaated seagass communities: a study of satial and temoal vaiation using tetheing techniques. Jounal of Exeimental Maine Biology and Ecology 107: Jablonski, D The toics as a souce of evolutionay velty though geological time. Natue 364: Jeanne, R. L A latitudinal gadient in ates of ant edation. Ecology 60: Julkunen-Tiitto, R Phelic constituents in the leaves of then willows: methods fo the analysis of cetain helics. Jounal of Agicultual Food Chemisty 33: Kubanek, J., M. E. Hay, P. J. Bown, N. Lindquist, and W. Fenical Ligid chemical defenses in the feshwate macohyte Sauuus cenuus. Chemoecology 11:1 8. Levin, D. A Alkaloid-beaing lants: an ecogahical esective. Ameican Natualist 110: Louda, S. M Distibution ecology: vaiation in lant ecuitment ove a gadient in elation to insect seed edation. Ecological Mogahs 52: MacAthu, R. H Geogahical ecology: attens in the distibution of secies. Hae and Row, New Yok, New Yok, USA. McNeill, S., and T. R. E. Southwood The ole of nitogen in the develoment of insect lant elationshis. Pages in J. B. Habone, edito. Biochemical asects of lant and animal coevolution. Academic Pess, London, UK. Mithen, R. A., F. Raybould, and A. Giamoustais Divegent selection fo seconday metabolites between wild oulation of Bassica oleacea and its imlications fo lant hebivoe inteactions. Heedity 75: Newell, S. Y., and D. Pote Micobial seconday oduction fom salt mash-gass shoots, and its kwn and otential fates. Pages in M. P. Weinstein and D. A. Keege, editos. Concets and contovesies in tidal mash ecology, Kluwe Academic Publishes, Dodecht, The Nethelands. Pennings, S. C., and M. D. Betness Using latitudinal vaiation to examine effects of climate on coastal salt mash atten and ocess. Cuent Toics in Wetland Biogeochemisty 3: Pennings, S. C., and M. D. Betness Salt mash communities. Pages in M. D. Betness, S. D. Gaines, and M. E. Hay, editos. Maine Community Ecology. Sinaue Associates, Sundeland, Massachusetts, USA. Pennings, S. C., T. H. Caefoot, E. L. Siska, M. E. Chase, and T. A. Page Feeding efeences of a genealist salt mash cab: elative imotance of multile lant taits. Ecology 79: Pennings, S. C., E. L. Siska, and M. D. Betness Latitudinal diffeences in lant alatability in Atlantic Coast salt mashes. Ecology 82: Rietsma, C. S., I. Valiela, and R. Buschsbaum Detital chemisty, gowth, and food choice in the salt-mash snail (Melamus bidentatus). Ecology 69: Rodiguez, E Ecogeogahic distibution of seconday constituents in Pathenium cifeae (Comositae). Biochemical Systematics and Ecology 5: Steinbeg, P. D Geogahical vaiation in the inteaction between maine hebivoes and bown algal seconday metabolites. Pages in V. J. Paul, edito. Ecological oles of maine natual oducts. Comstock Publishing Associates, Ithaca, New Yok, USA. Steinbeg, P. D., K. Edyvane, R. de Nys, R. Bidsey, and I. A. van Altena Lack of avoidance of helic-ich bown algae by toical hebivoous fishes. Maine Biology 109: Steinbeg, P. D., J. A. Estes, and F. C. Winte Evolutionay consequences of food chain length in kel foest communities. Poceedings of the National Academy of Sciences (USA) 92: Steinbeg, P. D., and I. van Altena Toleance of maine invetebate hebivoes to bown algal hlootannins in temeate Austalasia. Ecological Mogahs 62: Stiven, A. E., and S. A. Gadne Poulation ocesses in the ibbed mussel Geukensia demissa (Dillwyn) in a Noth Caolina salt mash tidal gadient: satial atten, edation, gowth and motality. Jounal of Exeimental Maine Biology and Ecology 160: Swihat, R. K., J. P. Byant, and L. Newton Latitudinal attens in consumtion of woody lants by swshoe haes in the easten United States. Oikos 70: Tagett, N. M., L. D. Coen, A. A. Boettche, and C. E. Tanne Biogeogahic comaison of maine algal olyhelics: evidence against a latitudinal tend. Oecologia 89: Valentine, J. F., and K. L. Heck, J The ole of leaf nitogen content in detemining tutlegass (Thalassia testudinum) gazing by a genealized hebivoe in the theasten Gulf of Mexico. Jounal of Exeimental Maine Biology and Ecology 258: Valiela, I., L. Koumjian, T. Swain, J. M. Teal, and J. E. Hobbie Cinnamic acid inhibition of detitus feeding. Natue 280: Van Alstyne, K. L., and V. J. Paul The biogeogahy of olyhelic comounds in maine macoalgae: temeate bown algal defenses dete feeding by toical hebivoous fishes. Oecologia 84: Vemeij, G. J Biogeogahy and adatation. Havad Univesity Pess, Cambidge, Massachusetts, USA. White, T. C. R The availability of invetebate hebivoes in elation to the availability of nitogen in stessed food lants. Oecologia 63: White, T. C. R The inadequate envionment: nitogen and the abundance of animals. Singe-Velag, Belin, Gemany. Zimme, M Sufactants in the gut fluids of Pocellio scabe (Isooda: Oniscidea), and thei inteactions with helics. Jounal of Insect Physiology 43: