B6 PRINCIPES [Vor. 35

Principes,35(2), 1991, pp. 86 93

EcologicalAspects of the Interaction between Chamaedoreatepejilote, a DioeciousPalm and Calyptocephalamarginipennis, a Herbivorous Beetle, in a Mexican Rain Forest

K. Oveue.* ANDR. DIRZo Centro d.eEcol.osia, WAM, Apartado Postal 7O-275, M6xico, 04510, D.F.

Assrnecr ers have made studies regarding intraspe- cific differencesin herbivory in the context Liebm. ex Mart., an tepejilote of polymorphic systemsin temperate abundant dioeciouspalm in the lowland rain forest of Los Tuxtlas (SoutheastMexico), typically bears evi- zones(e.g., Jones 1966, Cates1975, Dirzo dence of damage by Calyptocephala margini- and Harper I982a). Dioecy constitutes pennis Bohem. (Chrysomelidae).The beetles mark- one of the most common polymorphic sys- edly prefer this palm in comparison to other tems, particularly in some Neotropical on this site, including other sympatric speciesin the same . Additionally, beetlesshowed some mar- communities(Croat 1979, Bawa and Opler ginal preference for the foliage of female plants under 1975, Bawa 1980, Flores and Schemske experimental and field conditions. Sex-related phy- 1984, Bawa et al. 1985). Suggestionin tochemical attributes of the palm, such as secondary the literature of sex-baseddifferences in and nutritional characteristicsof the foli- comoounds herbivory include male inflorescencessuf- age, did not differ markedly between sexes and did not seemto be the proximal causesof the inter-sexual fering higher levels of predation than female differences in herbivory. Thus, the hypothesis that (Fryxell and Lulcefahr1967, differencesin resourceallocation to defenseand nutri Bawa and Opler 1978) and casesin which tional quality, resulting from sex-related differential herbivores preferentially feed either on allocation to reproduction, is not supported by this study. This investigation suggests that plant-herbi- leavesof female (Danell et al. 1985, Lovett vore interactionsin palms may be as complex as those Doust and Lovett Doust 1985) or male of other flowering plants. (Agren 1987) plants. However, folivory is extremely poorly known for dioecioustrop- Studieson the interactionsbetween her- ical plants, particularly palms. bivores and plants in tropical communities In the lowland rain forest of Los Tuxtlaso have been increasing markedly in recent SoutheastMexico, Charnaedoreatepej ilo- years (Coley 1983, Dirzo 1984, Janzen fe, a dioeciousunderstory palm, normally and Waterman 1984, Marquis 1984, bearsextremely variable levelsof leaf dam- among others). Also, much empirical evi- age and we explored the possibility that dencehas been obtainedwhich can be used this variation might explained, to some for the development of theories of plant- degreeat least, by inter-sexualdifferences. herbivore interactions (seeCrawley 1983, The plants on this site are normally eaten Rosenthaland Janzen 1979, Strong et al. by CaLyptocephala marginipennis, a l9B4). In this expandingfield, somework- chrysomelid beetle which appears to be largely monophagous.We measured the selectivity of this insect for the of

* Present address: Department of Botany, Faculty each sex in the field and laboratory. Also, of Science, Kyoto University, Kyoto, 606, Japan. we attempted to correlate the effect of L99ll OYAMA AND DIRZO: HERBIVORYCHAMAEDOREA herbivory with differences in contents of the leaves of male and female plants in the secondary compounds and nutritional field using two methods: a) an instanta- characteristicsof the sexes. neous measurement based on broad cat- egories of damage, and b) estimates of leaf area loss during a 3-month period. For the Materials and Methods first method, leaf damage was scored on an A to D scale (A < 25%; B > 25 < Study Species. The palm genus Cha- 5O%; C > 50 < 75%;D > 75%) n all maedorea includes approximately one the leaves of 48 male and 45 female plants. hundred speciesdistributed through Neo- The leaves of each plant were separated tropical forests from Brazil and Bolivia to into four age categories (I-IV). Contin- Guatemalaand the Central part of Mexico gency analyses (after a heterogeneity test (Standley and Steyermark 1958, Uhl and for the leaves of different age) were carried Dransfield 1987). All the in the out to assess the distribution of categories genus are dioecious (H. Quero, pers. of damage of the leaves of both sexes. b) comm.). In Los Tuxtlas, Mexico, C. tepe- For the second method, leaf damage was jilote is a common palm that forms dense also scored with a grid of 893 points on a patches in the forest understory where transparent plastic sheet (I0 cm x 5 cm). slight natural disturbanceshave occurred. This grid was placed on the adaxial surface The tallest individuals reach up to 5.0 m of the pinnae in a way that we could count in height. Plants of both sexes bear from the number of visible points coinciding with 3 to 5 leaves.The plants from Octo- ruptures of the lamina (as a product of ber through December. Both sexes have herbivore damage). We took care to check yellow green ; male flowers tend to all types of damage and differentiated them be slightly smaller (1.5 to 2.5 mm) than from the specific damage produced by C. females(2.5 to 3.5 mm). Single female marginipennis; this was, by far, the most plantscan bear from 5 to 400 (Oyama abundant type of damage (see Oyama 1984). Mortality of adult plants is due l9B4). We then calculated an index of mostly to branches or other debris falling leaf area removed as the quotient of the from the canopy. Tree falls frequently number of points registered (ruptured) to break apical meristems and kill the palms the total number of points of the grid (i.e., (Oyama1987, 1990). 893). Following this second method, we Eight speciesof Calyptocephala have recorded the cumulative damage on the been reported in Neotropical forests from leaves by marking a permanent area of Mexico to Brazil, paralleling the geo- observation (of equal size to the grid) on graphic distribution of the Chamaedorea the last two (youngest) leaves produced by species.The beetles are small (ca. 6 mm the plant. We recorded the damage long) with pale brown elitrae and a dark monthly for three consecutive months. For line at the margin. This insect feeds on this analysis, we chose randomly 15 indi- leavesof C. tepejilote making specificlon- viduals of each sex in three categories of gitudinal ruptures along the leaf veins, height (0 to 2.0 m; 2.0 to 3.5 m; > 3.5 which are easily recognizedin the field. m). Two-way analyses of variance were Study Site. This study was conducted done in function of sex and height (age) in a 600 m2 permanent plot in a lowland of the plants, after arcsin transformation rain forest in the Gulf of Mexico (95"14'- of the data. 95o09'W, lBo34'-18o36'N),in the bio- In a second line of investigation, intact logical reserve Estaci6n de Biologia Trop- pinnae of male and female plants of the ical Los Tuxtlas at 150-650 m altitude. same size and age were bagged together Methods. We recorded herbivory on and exposed to 6 beetles previously starved B8 PRINCIPES [Vor-.35

of m'ale Table 1. Contingency analysis for the number of leaaes of dffirent age _(M) in each category do'ftiage. Values and.female g1 ptanis o/Ciramaedorea tepejilote -of in pJrenthese. b"io* each pair of values are the absolutedeviations from those expected if beetlesdamaged leaves regardless of sex; the sign of the deviation is shownin parentheses after each ,r"G. Chi-tqnare analysesare shown for leaf age, the pooled data of leaves .d@

Category of beetle damage Leaf age Sex . Total X2 I M 90576 6 924 6.96 5 n.s. F 1,045 l5 I 6 r,067 .ll II M 827 46 13 897 t.29 3 n.s. F 929 60 II 14 1,014 il1 M 7rr (+) 72(-) 22(+) 19(-) 424 26.88 <0.001 F 690(-) 147(+) 2r (-) 34 (+) 892 (38) (33) (l) (6) M 470(+) r22 (-) 40 (-) 28 (-) 660 36.76 <0.001 F 314(-) I53(+) 43(+) 56 (+) s66 (48) (26) (5) (t 7) Total 71.89 I2 <0.001 Pooled M 2,e13(-) 247(-) Bl (+) 64 (-) 3,30s 31.07 3 F 2,978(+) 375(+) 76(-) 1r0 (+) 3,700 (6s) (r28) (s) (46) Heterogeneity 40.42 T2 <0.001

for 24 hours. The experiments liverecar- Results ried out in a laboratory arranged for this purpose. We recorded the leaf area con- In the survey of leaf damagescored on iumed for each type (sex) of plant 12 and a visual scale, the heterogeneity test 24 hours after. Each experiment consisted between leaves of different age was sig- : oI six replicates.Statistical analyses (t-tests) nifrcant (X" 40.82:.P < 0.001) and the were done to compare the mean leaf area resultsof eachleaf agewere examinedinde- eaten of male and female plants. All the pendently (Table l). The heterogeneity statisticalanalyses were done following Zar appears to be due to the fact that leaves (1974). in categories I and II did not show sex- A third line of study consistedof a phy- related differenceswhile categoriesIII and tochemical screening of secondary com- IV were the only caseswhere sexesdiffer pounds (alkaloids, tannins, flavonoids, sa- in the levels of damage. Moreover, the ponins and cyanogenic compounds) and leavesof the youngestcategory were mostly nutritional characteristics (protein, fiber, undamaged (905/924 for female and sugar, mineralsand fat contents)as Pos- I,045/L,067 for maleplants) and damage sible factors affecting the intensity of her- gradually increases with leaf age. Thus, bivory. Young and mature leaves of each sex-related differences could only be sex were collected, dried and analyzed in deteetedwhen damagewas high in general. the laboratory (see Dominguez(1979) for For leaves of age categories III and IV, a description of techniquesfor analysesof clearly females were over-representedin secondarycompounds, and Flores(I977) the heaviest category of damage (D) and for a descriptionof techniquesfor assessing in category B, while males were under- nutritional characteristics). representedin these categories.Males, in r99rl OYAMA AND DIRZO: HERBIVORYCHAMAEDOREA contrast, were over-represented in the o- (A) lightest categoryof damage and females J.F under-represented(Table l). o The time-courseof damagefor male and It female plants of the three categories of o size(Fig. I ) showedthat leaveswith herbiv- o ory were high in the young (smallest)plants and decreasedwith plant age. An inter- E 0 sexual difference was only apparent in the o smallest plants after 3 months. The sta- tistical analysisrevealed that the effect of plant sex was not significant(t' : I.4l; P > 0.05) while plant size significantly affectedthe level of damage(F : 5.01; P < 0.05). The interaction size X sex was also not statistically significant (F : O.82:P > 0.05). The feedingtrials carried out under con- io trolled conditionsshowed a significantpref- 0 o erence for female tissue when recordings E : o were made 12 hours after (t 3.017; P 0 o < 0.01, Table 2). However,the intensity of the preferencedecreased when the bee- 6 tles were maintained within the bagged ei plantsfor 24 hours(t: 2.367;P > 0.05), I ^i suggestingthat the proximal factors deter- 6 mining the relative acceptability reduce their influence with time. There was not any preference for male plants. The chemical screening of secondary metabolites showed the presence of fla- vonoids and tannins in the young and mature leaves of plants of all three size categoriesof both sexes(Table 3a). Alka- loids, saponinsand cyanogeniccompounds were not detectedin any plant sizeof either sex. The nutritional analysisshowed some qualitative differences in fiber, fat, min- erals and suear contents between sexes. E

d

I o

I. Time-course of the cumulative leaf damage between male (O) and female (O) plants oI Chamae- dorea tepejilote, by Calyptocephala marginipen- zis. Values are means of + I S.D. the index of leaf damage. A to C represent individuals of diferent height. (The index, when multiplied by I00, gives an estimate of the percent leaf area damaged per plant; details about the index. see text.) PRINCIPES lVoL. 35

\ I

2. Leaf ol showing heavy damage by the beetle Calyptocephala marginipennis.

The values of the first three were slightly Discussion greater in the leaves of male plants and the latter was greater in female plants The recognizable damage of C. mar- (Table 3b). Unfortunately, these chemical ginipennis on the foliage of C. tepejilote data are not amenable to statistical anal- in the field is an indication of the Prefer- yses,but the differencesappear to be only ence of the herbivore for this particular marginal. plant. Moreover, preliminary laboratory studies(Oyama 1984) strongly suggestthat Table 2. Comparison of feeding pref- of the five sympatric specieso{ Chatnae' erences o/ Calyptocephala marginipennis dorea in Los Tuxtlas, C. tepejilote is, by mar- for the leaues of male (M) and female far, the most acceptablefood for C. (F) plants o/Chamaedorea tepejilote. n ginipennis. On the other hand, these i represents the number of assaysand the observationsand laboratory results are not valuesare means+ I S.D. only indicative of the marked degree of I preference on the part of the herbivore, Leaf area removed (mm 2) but also suggesta tendency for exclusivity After 12 h After 24 h of this herbivore damage being experi- (n:24) (n:24) enced by the plant. The mature leaves of M 4.08+ 2.58 6.27+ 3.73 C. tepejilote accumulate the beetles'dam- F 7.03+ 3.30 9.40+ 5.24 age for severalmonths and when combined t 3.O2 2.37 with the fall of branches or other debris P <0.05 n.s. from the canopy constitute a major cause l99ll OYAMA AND DIRZO: HERBIVORYCHAMAEDOREA 9L

Table 3. Chemical attributes of leaoes of male (M) and fenale (F) plants of Cha- maedorea tepeiilote a. Secondary compounds* Cyanogenic Sex Leaf age Alkaloids Saponins Tannins Flavonoids capacity

M Young + TT Mature T TT F Yormg -f TT Mature T TT b. Nutritional components(7o) Carbo- Sex Proteins Fibers hydrates Minerals Total

M 20.56 A.I I 3l.lI 32.42 12.t4 100 F 20.55 3.53 27.41 37.70 |'0.81 r00 *-:absent;+:presenu ++ a more intense response.

of leaf death (Oyama t9B7). Experiments basis of differences in resource allocation of artificial defoliation showed detrimental betweenmale and female plants, in a sim- effects on the growth and reproduction ilar way to that some authors have sug- behavior of this plant (Oyama 1987, gestedfor other polymorphicsystems (Cates Oyama and Mendoza,1990). 1975; Dirzo and Harper 1982a, b). In Another issue of this paper relates to these studies, there is an indication that the suggestiveresults for a marginal female one of the morphs, the one preferred by tissue preference. Though in the labora- herbivores, displays also a greater com- tory there was consistently an initial pref- petitive ability, growth potential, or repro- erence for female tissue and in no case duction, presumably as a result of lower was male tissue preferably taken, the field investment of resources in defenses;the oodefended" data are not as conclusivealthough a sim- morph, instead, has fewer ilar tendency was observed for leaves of resourcesallocated to reproduction,growth intermediate to old age. The elucidation of or competitive ability. In C. tepeiilote, the role of other componentsof the natural female plants produce a substantiallymore history of the system (such as the role of costly reproductive structure (Oyama beetles' own predators and parasitoids)is 1987, Oyama and Dirzo l9B8): large, still obscure, but the results of this study heavy inflorescenceswhich then become (particularly those of the laboratory) war- heavy infructescencesthat remain attached rant further study in more detail and of to the plant for one year; male plants, in longer duration. contrast, producelong, thin and ephemeral The screeningof chemical components inflorescences which represent a much did not reveal any significant differences lower proportion of the biomassof the plant. betweensexes in the secondarymetabolites These differences are, in principle, com- and nutritional characteristicsand thus did patible with an interpretation on the not give any hint as to what could be the grounds of differential resource allocation, proximate cause behind the marginal dif- which might lead to greater susceptibility ferential damagesuffered by female plants. to herbivory by female plants. A testable explanation for the differ- The documented studies of folivory on ential herbivory can be proposed on the dioecious plants show contrasting results. 92 PRINCIPES lVoL. 35

Lovett Doust and Lovett Doust (1985) inflorescencesof Simarouba glauca earen less inflorescences? Evolution 32: found that female plants of Rumex ace- than staminate 673-676. tosella suffer heavier damage than male Cerrs, R. G. 1975. The interface between slugs plants and they hypothesizedthat female and wild ginger: someevolutionary aspects.Ecol- plants.with a lower leaf area, are in some ogy 56: 391-400. way more capable of withstanding envi- Cornv, P. D. 1983. Herbivory and defensivechar- of tree species in a lowland tropical In acteristics ronmental buffeting than male plants. forest. Ecol. Monog. 53: 209-233. contrast, Agren (1987) found that male Cnewrono, R. M. M. AND J. BALFOUR. 1983. leavesof Rubus channaernoruswere more Female-predominant sex ratios and ecophysio- damaged that female leaves. Different logical differentiation in arctic willows. J. Ecol. physiologicalresponses of each sex in dioe- 7l: 149-160. CRAvLEY, M. L. 1983. The dynamics of animal- cious species have been documented plant interactions. Studies in Ecology V. I0. (Crawford and Balfour 1983, Fox and Blackwell Scientific Pub., London. Harrison I98I, Zinmerman and Lechowi- Cnoer, T. B. 1979. The sexuality of the Barro cz 1982) but they appear to be due to Colorado Island flora (Panama). Phytologia 42: general pat- 319-348. local conditions rather than a DANELL, K., T. Eruqvrsr, L. EntcsoN, AND A. tern (Willson 1983). SetoltoNson. 1985. Sexuality in willows and The present paper provides some basic preference by bark-eating voles: defenseor not? aspectsof the natural history of the inter- Oikos 44: 82-90. Herbivory: a phytocentric over- action between C. tepejil.ote and,C. mar- DIRzo, R. 1984. view. 1n.'R. Dirzo and J. Sarukhan (eds.). Per- ginipennis; it exposesthe complexity of spectives on plant population ecology. Sinauer the system and suggeststhat some species AssociatesInc. Pub., Srmderland,pp. l4I-165. (pcpulations)of palms may be engagedin exo J. L. HARPER. 1982a. Experimental as complex and exciting interactions with studies on slug-plant interactions. III. Differ- in acceptability of individual plants of reported ences the herbivores as those commonly Trifolium repens ro slugs and snails. J. Ecol. for other flowering plants. 70: 10I 117. AND-. I9B2b. Experimental stud- ies on slug-plant interaction. IV. The perfor- Acknowledgments mance of cyanogenic and acyanogenic morphs in the field. J. Ecol. 70: Il9 138. We thank S. Sinaca and S. Guzm6n for Dourucunz, X. L. 1979. M6todosde Investigaci6n Fitoouimica.Limusa. M6xico. their field assistance,and S. Zaragoza for FLoRES,Vt. J. tgll . Bromatologiaanimal. Limusa, the identification of the insect. The study M6xico. waspartly supportedby a CONACyT grant FLoRES,S. AND D. W. Scnrusrn. 1984. Dioecy to R.D. and monoecy in the flora of Puerto Rico and the Virgin Islands: ecological correlates. Biotro- oica l6: I32-139. Fox, J. F. AND A. T. HARRISoN. 1981. Habitat Lrrnnerunn CIrno assortmentof sexesand water balance in a dioe- cious grass Hesperochloa kingii. Oecologia AcnnN, J. 1987. Intersexual diferences in phe- (Berl.) 49: 233-235. nology and damageby herbivoresand pathogens FRyxELL,P.A.er.rnM.J. Lurnresn. 1967. Ham- in dioeciousRubus chamaemorus L. Oecologia pea Schlecht.: possibleprimary host of the cot- (Berl.)72: l6l-169. ton boll weevil.Science 155: 1568-1569. BAwA, K. S. 1980. Evolution of dioecy in flowering JeuzrN, D. H. AND P. G. WerrnueN. 1984. A plants.Ann. Rev. Ecol. Syst. lI: 15-39. seasonalcensus of phenolics, fiber and alkaloids -, s. H. Burrocr, D. R. PERRY,R. E. COvrrrn, in foliage of forest trees in Costa Rica: some e.Nl M. H. Gnevuru. 1985. Reproductive biol- factors influencing their distribution and relation ogy of tropical rain forest trees. II. Pollination to host selectionby Sphingidaeand Saturniidae. systems.Amer. J. Bot.72:346 356. Biol. J. Linn. Soc. 2l: 439-454. eNo P. A. Oprnn. 1975. Dioecism in Joxns, D. A. 1966. On the polymorphism of cya- tropical forest trees. Evolution 29: 167-179. nogenesisin Lotus corniculatus L. I. Selection AND _. I9Tg. Whv are pistillate by animals. Can. J. Genet. Cytol. 8: 556-567. r99ll OYAMA AND DIRZO: HERBIVORYCHAMAEDOREA 93

Lovrrr Dousr, J. ano L. LovErr Dousr. 1985. liation on growth, reproduction and survival oI Sex ratios, clonal growth and herbivory in Rumer a Neotropical dioecious palm Chamaedorea acetosella. 1n; J. White (ed.). Studies on plant tepejilote. Biotropica 22: I 19-123. demography.Academic Press, London. pp.327- RoSENTHAL,G. A. ar.lnD. H. JeNzrN, EDS. 1979. 34r. Herbivores: their interaction with secondary plant Me,nqurs, R. J. 1984. Herbivory as a selective metabolites. Academic Press, New York. Iorce in Piper arieianum C. DC. (Piperaceae). STANDLEY,P. C. AND J. A. STEYERMARK.1958. Ph.D. Thesis, University of lowa. Flora ofGuatemala. Fieldiana Botany 24:216- Overr.re,K. 1984. Biologia comparativa entre indi- 25r. viduos masculinosy femeninosde Chamaedorea Srnoric, D. R., J. H. lawroN, ANDS.R. SourHwooD. tepejiloteLiebm. B.Sc. Thesis,Facultad de Cien- 1984. Insects on plants. Commmity Patterns cias. UNAM. M5xico. and Mechanisms.Blackwell Scientific Pub., Lon' 1987. Demografiay dinimica poblacional don. de Chamaedorea tepejilote Liebm. en la selva Unr, N. W. AND J. Dn-lxsrIrlo. 1987. Genera de Los Tuxtlas, Veracruz (M6xico). M.Sc. The- Palmarum. Allen Press, Lawrence, Kansas. sis, Facultad de Ciencias, UNAM, M6xico. WILrsoN, M. 1983. Plant reproductive ecology. 1990. Variation in growth and reproduc' John Wiley & Sons, New York. tion in a Neotropical dioecious palm, Chamae- Zen, J. H. 1974. Biostatistical analysis. Prentice' dorea tepejilote.J. Ecol. 78:648-663. Hall Inc., New Jersey. eNo R. Dmzo. 1988. Biomassallocation ZTMMERMAN,J. K. eNo M. J. LEcHovIcz. 1982. in the dioecious tropical palm Chamaedorea Resoonsesto moisture stressin male and female tepejilote and its life history consequences.Pl. plants of Rumex acetosello L. (Polygonaceae). Sp. Biol. 3:27-33. Oecologia(Berl.) 53: 305-309. ANDA. MENDoze. 1990. Effects of defo-

CALLINGALL EUROPEANS!

Arurouncing the formation of The European Palm Society, based in London, and catering to palm enthusiastsall over the European continent. "Chamaerops" containscolor and half-toneillustrations, together Our quarterly magazine, .We with articles by both amateur and professionalpalm growers. also attempt to cover members' wider growing interests, and include other exotic plants. Annual subscriptionis f,IO (!12 non E.E.C.). For more information pleasewrite to THn EunoprAN PALMSoclnrY, ToThe Palm Centre,563 Upper RichmondRoad, London SWI4 7ED, U.K.