Yanoviak: Treehole macroinvertebrates 265 CONTAINER COLOR AND LOCATION AFFECT MACROINVERTEBRATE COMMUNITY STRUCTURE IN ARTIFICIAL TREEHOLES IN PANAMA S TEPHEN P. Y ANOVIAK Department of Zoology, University of Oklahoma, Norman, OK 73019-0235 USA Current address: Evergreen State College, Lab 1, Olympia, WA 98505 A BSTRACT I investigated the effects of habitat color and location on community structure in artificial water-filled treeholes in the forest of Barro Colorado Island, Panama. The macroinverte- brate fauna of 9 replications (5 in understory, 4 in tree-fall gaps) of black, blue, red, and green 650 ml plastic cups were censused weekly for 7 wks. Macroinvertebrate abundance and species richness were greater in understory cups than in gap cups. Seven species colo- nized black cups exclusively. Black cups in the understory, and red and black cups in gaps, attracted more species on average than other colors. Species richness and abundance were consistently lowest in green cups. Species were more broadly distributed among cup colors in the understory, suggesting that diffuse light conditions influenced color perception. There was no overlap in species composition between water in the artificial treeholes and water held by red Heliconia bracts or green tank bromeliads. Key Words: color, microcosm, mosquitoes, phytotelmata, tree-fall gaps, tropics R ESUMEN Investigu? los efectos del color de h?bitat y localidad en la estructura de la comunidad en huecos de ?rboles artificiales llenos de agua en el bosque de Isla Barro Colorado, Panam?. La fauna macro invertebrada de 9 replicaciones (5 en el ?understory?, 4 en espacios abiertos de ?rboles ca?dos) de vasos pl?sticos de 650 ml de color negro, azul, rojo, y verde fueron obser- vadas semanalmente por 7 semanas. La abundancia macro invertebrada y la riqueza de es- pecies fueron mayor en vasos del ?understory? que en los vasos de espacios abiertos. Siete especies colonizaron vasos negros exclusivamente. Los vasos negros en el ?understory?, y los vasos rojos y negros en los espacios, atrajeron un promedio mayor de especies que otros co- lores. Riqueza y abundancia de especies fueron mas bajas consistentemente en vasos verdes. Las especies fueron distribuidas mas ampliamente entre colores de vaso en el ?understory?, sugiriendo que condiciones de luz difusa influyen percepci?n de color. No hubo ?rea com?n en composici?n de especies entre agua en los huecos de ?rbol artificiales y agua retenida por br?cteas rojas de Heliconia o bromelias de tanque verde. Container-breeding mosquitoes use a variety of physical and chemical cues when selecting ovi- position substrates (Bates 1949; Bentley & Day 1989), and many species show preferences for specific habitat colors (Frank 1985, 1986, and ref- erences therein). Most investigations of mosquito response to habitat color have been lab-based, and color preferences have not been studied for other taxa that typically coexist with mosquitoes in the field. Although mosquitoes generally domi- nate the macrofauna of tropical phytotelmata, several other insect taxa are also common in these habitats (e.g., Fish 1983). The effects of hab- itat color on colonization by non-mosquito aquatic macroinvertebrates in tropical phytotelmata are not known. Water-filled treeholes are common phytotel- mata in many temperate and tropical forests (Snow 1949; Kitching 1971), and occur in both tree-fall gaps and forest understory. Artificial treeholes (plastic containers containing leaf litter and rain water) are often used for field-based pop- ulation and community-level studies. These con- tainers provide suitable mimics of the treehole habitat, and typically attract the same fauna found in the natural setting (e.g., Fincke et al. 1997; Yanoviak, in press). Artificial treeholes are usually clear (Pimm & Kitching 1987; Srivastava & Lawton 1998), black or brown (Fincke et al. 1997; Yanoviak 1999a), or blue (O. M. Fincke, Dept. Zoology, Univ. Oklahoma, pers. comm.). How such color differences may affect community structure in these experimental systems has never been investigated. The potential importance of habitat color ex- tends beyond artificial treeholes; there is consid- erable variation in color among natural phyto- telmata. Common phytotelmata in the lowland forests of Panama include water-filled treeholes, tank bromeliads (e.g., Vriesia spp., Guzmania spp.), fallen palm petioles and spathes (e.g., of As- trocaryum standleyanum Bailey), and exocarps of 266 Florida Entomologist 84(2) June 2001 fallen Tontelea ovalifolia (Miers) A. C. Smith fruits. Treehole interiors typically appear black or brown; bromeliads are generally green; palm spathes and T. ovalifolia husks are initially cream-colored, but gradually become orange, red- dish, and eventually dark brown as they age (pers. obs.). In addition, the red and green bracts of Heliconia spp. often contain sufficient water to support aquatic macroinvertebrate assemblages (e.g., Seifert & Seifert 1979; Naeem 1988; Louni- bos & Machado-Allison 1993). This study examined the effects of container color and location (tree-fall gap or forest under- story) on colonization of artificial treeholes by macroinvertebrates under field conditions. Spe- cifically, I hypothesized that container location and color would affect macroinvertebrate species composition, species richness, and abundance. Earlier work on this system showed that highly exposed treeholes (in forest canopy) contained fewer species than holes in the understory (Yanoviak 1999b). Thus, I predicted that macroin- vertebrate species richness and abundance would be lower in gap holes of this study. Based on field observations and other studies showing oviposi- tion preference for dark containers (e.g., McDaniel et al. 1976), I predicted that black cups would attract more species than other colors. Fi- nally, because color is a potentially important cue for species colonizing bromeliads (Frank 1985, 1986), I expected that green cups would attract some taxa (e.g., Wyeomyia spp. mosquitoes) that normally inhabit tank bromeliads in Panama. M ATERIALS AND M ETHODS In mid-July 1997, I established 9 replications of four colored plastic cups (black, blue, green, and red; 650 ml, 8.5 cm diameter ? 12 cm height - Churchill Container Corp., Shawnee, KS, USA) in the seasonally wet tropical forest of Barro Colo- rado Island (BCI), Panama (see Leigh et al. 1996 for a site description). Four replicates were lo- cated in tree-fall gaps and 5 in forest understory. I secured each cup to a pole-sized (10-20 cm diam- eter) tree approximately 1 m above the ground (as described in Yanoviak 1999a). Cups initially con- tained ca. 400 ml rain water and a strip of balsa wood (0.2 ? 4 ? 16 cm) as an oviposition site for in- sect colonists (Novak & Peloquin 1981). The balsa strip was small enough relative to the cup that it had little effect on general color appearance of the habitat. Holes drilled in the cup rims maintained water levels ca. 100 ml below capacity. Thus, any insect entering to oviposit was exposed to the cup color from all sides and from below. Cups within a replication were separated by > 5 m and the dis- tance between replicates was > 100 m. I allowed rain water and detritus to accumulate naturally. The macroinvertebrate fauna (organisms > 0.5 mm) of all cups was censused weekly for 7 wks be- ginning 26 July 1997. I occasionally collected sub- samples of pupae and late-instar larvae to confirm field identifications. Water temperatures were measured at irregular intervals with a Corning? modular electronic probe. The experiment was terminated after 7 wks because detritus accumu- lations and fungal or algal growth changed the in- terior color of some cups to dark brown. In addition to the fauna of artificial treeholes, I qualitatively sampled macroinvertebrates living in nearby water held by 14 tank bromeliads, 20 T. ovalifolia exocarps, 12 fallen palm spathes, and 12 Heliconia latispatha Benth. bracts (from 12 different plants) for comparison with the species composition of communities in the plastic cups. Taxa that could not be identified in the field were collected and reared in the lab. Differences in macroinvertebrate species rich- ness and abundance in the artificial tree holes were analyzed with 2-way repeated-measures ANOVAs using location (gap or understory) and color as main effects. When significantly differ- ent, means were compared with Ryan-Einot-Gab- riel-Welsch multiple range tests (SAS 1989). Abundance data were log(x+1) transformed prior to analysis to correct variance heterogeneity (Sokal & Rohlf 1981), but all means presented in results were calculated from untransformed data. S?rensen?s (1948) coefficient of similarity [C s = 2j ? (a + b) -1 , where j = number of species common to two treatments, and a and b = the number of spe- cies in each treatment] was used to quantify over- lap in species composition (all replicates pooled within gap or understory locations) among the four cup colors. Reflectance data for the red, blue, and green cups were obtained with an Ocean Optics? S2000 spectrophotometer. Measurements were taken under dim illumination to reduce the variance at- tributable to environmental lighting. The probe was dark-spectrum calibrated by placing it into a black plastic bag, and a reference spectrum was obtained by measuring an Ocean Optics reflec- tance standard (a white piece of plastic). Three separate recordings, each resulting in 1570 data points between 360 and 860 nm, were taken from each cup. Reflectance spectra for the red, blue, and green cups are shown in Fig. 1. R ESULTS Artificial treeholes located in the understory were colonized by more individuals and more spe- cies than cups in tree-fall gaps (Tables 1 and 2; compare Figs. 2 and 3). Conditions in the cups dif- fered dramatically between locations. Tempera- tures in gap cups exceeded 40 ? C on sunny days, and all gap cups contained abundant filamentous algae by Week 5. In understory cups, tempera- tures never exceeded 32 ? C and algae were virtu- ally nonexistent. There were no significant inter- Yanoviak: Treehole macroinvertebrates 267 actions between treehole color and location for both macroinvertebrate species richness and abundance (Table 1), so effects of color were tested separately for gap and understory locations. Average species richness and abundance dif- fered among cup colors in both gaps and under- story (Table 1), but significant effects did not occur in either location until Week 4 (based on Ryan post-hoc tests). Species richness was great- est in black cups in the understory (Fig. 2a) and black and red cups in gaps (Fig. 3a) after 7 wks of colonization. Macroinvertebrate abundance was significantly lower in green cups in forest under- story, but did not differ among black, blue, and red cups on Weeks 6 and 7 (Fig. 2b). Black and red cups in gaps contained more individuals than did green cups by the end of the study (Fig. 3b). The significant time*treatment interaction for under- story species richness, but not for abundance (Ta- ble 1), reflects the early successional status of the artificial treeholes (Yanoviak, in press). Species richness in understory cups increased over time whereas abundance remained relatively constant (Fig. 2). The lack of a similar time interaction in gaps is attributed to high variance (Fig. 3). Black cups were colonized by the most taxa overall, whereas green cups attracted the fewest species (Table 2). Seven species, including four non-mosquito taxa, occurred exclusively in black cups. Overlap in species composition among dif- ferent colors was marginally greater in the under- story than in gaps (Wilcoxon 2-sample test using S?rensen?s coefficients obtained for each location, P = 0.077), and assemblages in understory black cups showed higher similarity to assemblages in red and blue cups than to those in green cups (Ta- ble 3). Several species that avoided blue and/or green cups in gaps colonized one or both of those colors in the understory (Table 2). There was no overlap between the species com- position of water held by bromeliads or H . latis- patha bracts on BCI and the cups used in this experiment. However, several treehole species were found in water held by palm spathes and T. ovalifolia exocarps; both were dominated by the mosquitoes Trichoprosopon digitatum (Rondani) and Limatus spp. D ISCUSSION My results show that the location and color of artificial treeholes can influence species richness, abundance, and composition in container commu- nities. As predicted, macroinvertebrate species richness and abundance were lower in artificial treeholes located in gaps. This result is likely due to oviposition preferences for shaded vs. exposed sites. Prior work on this system showed that some species are more common in shaded holes (Yanoviak 1999c) and that highly exposed canopy tree holes contain fewer species than understory holes (Yanoviak 1999b). It is also probable that extremely high water temperatures in gap cups depressed richness and abundance through larval mortality. Larvae of several mosquito species can- not survive temperatures above 40 ? C for more than a few minutes (Bates 1949). Finally, abun- dant algae in gap cups may have deterred ovipo- sition by chemical or other means, and occa- sionally caused larval mortality by restricting their movements within the cups (pers. obs.). Fig. 1. Reflectance spectra for the red, blue, and green cups used in this study. Each point is the mean of 30 data associated with 10 wavelength measurements (3 replicates ? 10 adjacent wavelengths), and each mean was plotted against its median wavelength value. Vari- ance was very low; error bars (? 1 SD) are shown only for a subset of the means. T ABLE 1. R EPEATED - MEASURES ANOVA OUTPUT . SS = TYPE III SUM OF SQUARES . T IME = P- VALUE FOR UNIVARIATE TEST OF TIME * TREATMENT INTERACTION . * = P < 0.05, ** = P < 0.005. Treatment Abundance Richness SS df F Time SS df F Time Location 10.45 1,28 16.83** 0.71 31.74 1,28 26.07** 0.02 Color 16.21 3,28 8.70** 0.07 66.94 3,28 18.32** <0.01 Location*Color 0.36 3,28 0.20 0.68 4.21 3,28 1.15 0.13 Color, Gap 5.64 3,12 7.72** 0.19 17.21 3,12 11.61** 0.40 Color, Underst. 11.59 3,16 4.27* 0.35 58.54 3,16 11.08** <0.01 268 Florida Entomologist 84(2) June 2001 Cup color significantly affected macroinverte- brate species richness and abundance in both lo- cations and, as predicted, black cups attracted the most species overall. Although many taxa colo- nized more than one color of cup, the general dis- tribution of species among colors suggests that most prefer to oviposit in dark containers or avoid green oviposition sites. This pattern is well docu- mented for several mosquito species (e.g., Will- iams 1962; Wilton 1968; McDaniel et al. 1976; Hilburn et al. 1983; Beehler et al. 1992; Jones & Schreiber 1994; also reviewed by Frank 1985), and at least partly explains why the green cups of this study contained few species and individuals. Differences in distributions of species among cup colors between gap and understory locations most likely reflect differences in how colors are perceived in these light environments. (One ex- ception among the taxa found in this study is the annelid Dero sp., which is primarily dispersed by phoresy.) Most insects are unable to see red (Chapman 1998), thus red containers probably appear dark gray to potential colonists. Wave- lengths of incident light in forest understory are shifted toward the blue/green end of the spectrum by the surrounding vegetation and toward red un- der overcast conditions (Endler 1993), thus blue and green cups in forest understory may also T ABLE 2. D ISTRIBUTION OF INVERTEBRATE TAXA AMONG DIFFERENT COLORED CUPS IN FOREST UNDERSTORY AND TREE - FALL GAP LOCATIONS . V ALUES ARE MAXIMUM NUMBER OF INDIVIDUALS OBSERVED IN A CUP WITHIN A TREAT- MENT . M INIMUM ABUNDANCE WAS ZERO IN MOST CASES . M EAN ABUNDANCE AND MEAN RICHNESS ARE THE AV- ERAGE ( ? 1 SD) NUMBER OF INDIVIDUALS AND SPECIES WITHIN A TREATMENT , ALL CENSUS DATES COMBINED . Taxon Understory Gap Black Blue Green Red Black Blue Green Red Annelida: Naididae Dero sp. 115 7 0 0 0 0 0 0 Odonata: Pseudostigmatidae Mecistogaster spp. 1 0 0 0 1 0 0 0 Coleoptera: Scirtidae Prionocyphon sp. 16 0 0 0 4 0 0 0 Diptera: Ceratopogonidae Bezzia snowi Lane 5 2 0 5 4 0 0 0 Forcipomyia spp. 26 0 0 1 3 0 37 0 Diptera: Chironomidae Chironomus sp. 0 0 0 0 1 0 0 0 Diptera: Culicidae Aedes terrens (Walker) complex 35 1 2 35 9 0 0 15 Anopheles eiseni Coq. 4 0 0 0 0 0 0 0 Culex conservator D. & K. 8 5 0 0 5 0 0 0 C. corrigani D. & K. 33 0 0 18 2 0 0 0 C. urichii (Coq.) 21 59 10 58 12 2 0 44 Haemagogus equinus Theobald 7 6 1 9 8 4 0 23 H. leucotaeniatus (Komp) 0 0 0 1 0 0 0 0 H. lucifer (Howard, Dyar & Knab) 0 0 0 15 0 9 0 0 Limatus assuleptus (Theobald) 16 15 15 38 6 5 1 32 Orthopodomyia fascipes (Coq.) 0 0 0 0 1 0 0 0 Toxorhynchites theobaldi (D. & K.) 1 0 0 0 1 0 0 0 Diptera: Psychodidae Telmatoscopus spp. 0 0 0 0 8 0 3 0 Diptera: Syrphidae Copestylum rafaelanum (Townsend) 69 0 0 0 0 0 0 0 Diptera: Tipulidae Sigmatomera spp. 0 0 0 0 0 0 5 0 Total no. of taxa represented 14 7 4 9 14 4 4 4 Mean abundance 22.3 (31.20) 10.5 (16.18) 3.0 (4.87) 19.1 (14.89) 4.8 (4.80) 1.3 (2.10) 2.1 (6.95) 11.3 (13.41) Mean richness 2.5 (1.71) 1.2 (1.10) 0.7 (0.77) 2.0 (1.17) 1.3 (1.00) 0.4 (0.56) 0.2 (0.55) 0.9 (0.80) Yanoviak: Treehole macroinvertebrates 269 have been perceived as gray or reddish. Results of this study support such a conclusion. Blue and red cups attracted more species overall in the un- derstory than in gaps, and species that colonized three or more colors tended to be more abundant in red and black. The mosquitoes Culex urichii (Coquillett) and Limatus assuleptus (Theobald) are exceptions to the latter, but both are also hab- itat generalists and will colonize water in almost any container (pers. obs.). Five of the 7 species found only in black cups occurred in very low abundance. The midge Chi- ronomus sp. and the mosquito Orthopodomyia fascipes (Coquillett) are normally more abundant in artificial treeholes than was observed here (Yanoviak, in press), so their distribution in this study may not reflect color preferences. Despite its low abundance, the presence of Anopheles eis- eni Coquillett only in black cups probably reflects a color preference; several other Anopheles spe- cies prefer to oviposit in dark containers (re- viewed by Frank 1985). Mecistogaster spp. and Toxorhynchites theobaldi (Dyar & Knab) are top predators in this system, and typically exist in low abundance due to competition and cannibal- ism (Fincke 1999). Treeholes are a limiting repro- ductive resource for pseudostigmatids (Fincke 1992a), and the species occurring on BCI avoid other phytotelmata, including bromeliads (Fincke 1992b). Lab studies showed that some Toxorhyn- chites species prefer to oviposit in black contain- ers (Hilburn et al. 1983; Jones & Schreiber 1994). Therefore, it is reasonable to conclude that the presence of these two taxa only in black cups re- flects a habitat color preference. Fig. 2. Mean (? SE) macroorganism species richness (A) and abundance (B) in different colored cups located in forest understory. N = 5 for each mean. Within a week, means followed by the same letter do not differ based on post-hoc tests (only noted where differences oc- curred; a single letter was used for overlapping means). Some error bars were omitted for clarity. Fig. 3. Mean (? SE) macroorganism species richness (A) and abundance (B) in different colored cups located in tree fall gaps. N = 4 for each mean. Within a week, means followed by the same letter do not differ based on post-hoc tests (only noted where differences occurred; a single letter was used for overlapping means). Some er- ror bars were omitted for clarity. TABLE 3. S?RENSEN?S SIMILARITY COEFFICIENTS FOR MACROINVERTEBRATE ASSEMBLAGES OCCURRING AMONG THE FOUR CUP COLORS IN UNDERSTORY AND TREE-FALL GAP LOCATIONS. Understory Gap Black Blue Red Black Blue Red Blue 0.667 ? ? 0.333 ? ? Red 0.609 0.625 ? 0.444 0.750 ? Green 0.444 0.727 0.615 0.333 0.250 0.250 270 Florida Entomologist 84(2) June 2001 The lack of species overlap between the colored cups and similarly colored phytotelmata on BCI can be explained by physical attributes of these systems other than color. First, the general form of a container may influence colonization. For ex- ample, Frank (1986) used artificial bromeliads to show that ovipositing mosquitoes are sensitive to the general shape of plant containers. Second, macroinvertebrate colonization is influenced by the vertical position of phytotelmata. Fallen fruit husks and palm spathes occur at ground level, whereas all of the containers used in this study were ca. 1 m above the ground. Some mosquito species in the BCI forest (e.g., Limatus spp., T. digitatum) are sensitive to this difference in height (Yanoviak 1999b, in press). Finally, many species respond to chemical cues when selecting oviposition sites (Bentley & Day 1989), and these cues may complement or exceed the effects of color (Lounibos & Machado-Allison 1993). CONCLUSIONS My results suggest that differential coloniza- tion and harsh environmental conditions in tree- fall gaps tend to reduce the abundance and diver- sity of macroinvertebrates in artificial tree holes. Container color appears to play a role in habitat selection for several species in this system, but perception of color may be confounded by variable light conditions in forest understory. Additional studies under controlled laboratory conditions are needed to determine oviposition preferences of the many non-mosquito colonists of tropical phytotelmata. ACKNOWLEDGMENTS A. Almanza assisted in the field and R. Fuller provided reflectance data for the cups. Comments from L. P. Lounibos, D. Srivastava, and an anony- mous reviewer improved the manuscript. This work was supported by funds from the IIE/Fulbright Foundation and the University of Oklahoma. LITERATURE CITED BATES, M. 1949. The natural history of mosquitoes. MacMillan, Gloucester, MA. BEEHLER, J., S. LOHR, AND G. DEFOLIART. 1992. Factors influencing oviposition in Aedes triseriatus (Diptera: Culicidae). Great Lakes Entomol. 25: 259-264. BENTLEY, M. D., AND J. F. DAY. 1989. Chemical ecology and behavioral aspects of mosquito oviposition. Annu. Rev. Entomol. 34: 401-421. CHAPMAN, R. F. 1998. The insects: structure and func- tion. 4th ed. Cambridge Univ. Press, Cambridge. ENDLER, J. A. 1993. The color of light in forests and its implications. Ecol. Monogr. 63: 1-27. FINCKE, O. M. 1992a. Interspecific competition for tree holes: consequences for mating systems and coexist- ence in neotropical damselflies. American Nat. 139: 80-101. FINCKE, O. M. 1992b. Behavioural ecology of the giant damselflies of Barro Colorado Island, Panama, pp. 102-113. In D. Quintero and A. Aiello (eds.). Insects of Panama and Mesoamerica. Oxford Univ. Press, Oxford. FINCKE, O. M. 1999. Organization of predator assem- blages in Neotropical tree holes: effects of abiotic fac- tors and priority. Ecol. Entomol. 24: 13-23. FINCKE, O. M., S. P. YANOVIAK, AND R. D. HANSCHU. 1997. Predation by odonates depresses mosquito abundance in water-filled tree holes in Panama. Oecologia 112: 244-253. FISH, D. 1983. Phytotelmata: flora and fauna, pp. 1-27. In J. H. Frank and L. P. Lounibos (eds.). Phytotel- mata: terrestrial plants as hosts for aquatic insect communities. Plexus, Medford, NJ. FRANK, J. H. 1985. Use of an artificial bromeliad to show the importance of color value in restricting the coloni- zation of bromeliads by Aedes aegypti and Culex quin- quefasciatus. J. Am. Mosq. Control Assoc. 1: 28-32. FRANK, J. H. 1986. Bromeliads as ovipositional sites for Wyeomyia mosquitoes: form and color influence be- havior. Florida Entomol. 69: 728-742. HILBURN, L. R., N. L. WILLIS, AND J. A. SEAWRIGHT. 1983. An analysis of preference in the color of ovipo- sition sites exhibited by female Toxorhynchites r. rutilus in the laboratory. Mosq. News 43: 302-306. JONES, C. J., AND E. T. SCHREIBER. 1994. Color and height affects oviposition site preferences of Toxo- rhynchites splendens and Toxorhynchites rutilus rutilus (Diptera: Culicidae) in the laboratory. Envi- ron. Entomol. 23: 130-135. KITCHING, R. L. 1971. An ecological study of water-filled tree-holes and their position in the woodland ecosys- tem. J. Anim. Ecol. 40: 281-302. LEIGH, E. G., JR., A. S. RAND, AND D. M. WINDSOR (eds.). 1996. The ecology of a tropical forest. 2nd ed. Smith- sonian, Washington, DC. LOUNIBOS, L. P., AND C. E. MACHADO-ALLISON. 1993. Field test of mosquito ovipositional cues from Vene- zuelan phytotelmata. Florida Entomol. 76: 593-599. MCDANIEL, I. N., M. D. BENTLEY, H.-P. LEE, AND M. YATAGAI. 1976. Effects of color and larval-produced oviposition attractants on oviposition of Aedes trise- riatus. Environ. Entomol. 5: 553-556. NAEEM, S. 1988. Predator-prey interactions and com- munity structure: chironomids, mosquitoes and copepods in Heliconia imbricata (Musaceae). Oecolo- gia 77: 202-209. NOVAK, R. J., AND J. J. PELOQUIN. 1981. A substrate modification for the oviposition trap used for detect- ing the presence of Aedes triseriatus. Mosq. News 41: 180-181. PIMM, S. L., AND R. L. KITCHING. 1987. The determi- nants of food chain lengths. Oikos 50: 302-307. SAS. 1989. SAS/STAT user?s guide, volume 2, GLM- VARCOMP. SAS Institute Inc., Cary, NC. SEIFERT, R. P., AND F. H. SEIFERT. 1979. A Heliconia in- sect community in a Venezuelan cloud forest. Ecol- ogy 60: 462-467. SNOW, W. E. 1949. The arthropoda of wet tree holes. Ph.D. Dissertation, Univ. of Illinois, Urbana. 235 p. SOKAL, R. R., AND F. J. ROHLF. 1981. Biometry. W. H. Freeman & Co., New York. S?RENSEN, T. 1948. A method of establishing groups of equal amplitude in plant sociology based on similar- ity of species content. K. Danske Vidensk. Selsk. Biol. Skrift 5: 1-34. Yanoviak: Treehole macroinvertebrates 271 SRIVASTAVA, D. S., AND J. H. LAWTON. 1998. Why more productive sites have more species: experimental test of theory using tree-hole communities. Ameri- can Nat. 152: 510-529. WILLIAMS, R. E. 1962. Effect of coloring oviposition me- dia with regard to the mosquito Aedes triseriatus (Say). J. Parasit. 48: 919-925. WILTON, D. P. 1968. Oviposition site selection by the tree-hole mosquito, Aedes triseriatus (Say). J. Med. Entomol. 5: 189-194. YANOVIAK, S. P. l999a. Effects of leaf litter species on macroinvertebrate community properties and mos- quito yield in Neotropical tree hole microcosms. Oecologia 120: 147-155. YANOVIAK, S. P. 1999b. Community structure in water- filled tree holes of Panama: effects of hole height and size. Selbyana 20: 106-115. YANOVIAK, S. P. 1999c. Distribution and abundance of Microvelia cavicola Polhemus (Heteroptera: Veli- idae) on Barro Colorado Island, Panama. J. New York Entomol. Soc. 107: 38-45. YANOVIAK, S. P. 2001. The macrofauna of water-filled tree holes on Barro Colorado Island, Panama. Bio- tropica 33: (in press).