In Local Ground Ant Assemblages of the Americas

ANALES DE BIOLOG~A,19 (Biología Animal, 8) 1993: 29-37 SECRETARIADO DE PUBLICACIONES - UNIVERSIDAD DE MURCIA

RELATIVE REPRESENTATION OF PHEIDOLE (HYMENOPTERA: FORiWCIDAE) IN LOCAL GROUND ANT ASSEMBLAGES OF THE AMERICAS

H. G. Fowler

Recibido: 10 junio 1993 Aceptado: 28 diciembre 1993 RESUMEN Participaci6n relativa de Pheidole (Hymenoptera: Fomdddae) en comunidadm locales de hormigas de suelo en las Américas

En 195 estudios publicados sobre comunidades de hormigas de suelo de las Américas, la representación faunística de Pheidole fue mayor que la de cualquier otro género. No se encontró una relación significativa entre el porcentaje de la representación faunística y la latitud, pero su representación fue mayor en los trópi- cos. En cultivos fmctíwlos se detectó un porcentaje mayor de especies de Pheidole que en otros tipos de hábitat. Adicionalmente, el porcentaje de especies de Pheidole relacionó negativamente con el porcentaje de especies de Formicinae. Este efecto fue mayor que la influencia de la latitud sobre Fomicinae. No se encontró una relación negativa significativa entre Pheidole y otros grupos taxonómicos. Este resultado sugiere que las especies de Pheidole pueden competir con las especies de Formicinae, o que estos dos grupos están ubicados en los extremos opuestos de un gradiente de estrés ambiental.

Palabras clave: Comunidad, hormiga, latitud, hábitat, patrón, diversidad, Pheidole. SUMMARY

Using 195 published ground ant wmmunity surveys from the Amencas, the fauna1 representation of Pheidole was greater than any other genus. No significant relationship between the percent faunal representation with latitude was found, although their representation was much richer in the tropics. Tree crops had significantly higher percentages of species of Pheidole than did other habitat types. Furthermore, a significant negative relationship was found between the percent representation of Pheidole and al1 formicines. This effect was stronger than the relationship between formicines and latitude. No significant negative relations were found for other subfamilies. This suggests that Pheidole may compete strongly with formicines, or that these groups are placed on opposite ends of an enviromental stress gradient. Key words: Community, ant, latitude, habitat, pattem, diversity, Pheidole.

Departamento de Ecologia, Instituto de Biiitncias, UNESP, 13506-900 Rio Claro, Seo Paulo, Bradl. Telepho- ne: 195-24-2588. Fax: 195-344433. 30 H. C.FOWLER ANWS DE MoLOGh, 19 (1993)

INTRODUCTION cies richness. Comparisons with other taxonomic levels were also perfomed likewise. Taxo- The cosmopolitan ant genus Pheidole has nomy follows CREIGmON (1950)s KEMPF apparently evolved and diversified since the ( 1972) and BROW ( 1973). Oligocene. It has replaced members of the Do- lichoderinae in the temperate regions and has diversified at high rates in the tropics (BROWN, RESULTS AND DISCUSSION 1973). In the Americas, the ant genus Pheidole is the most species rich (CRGGHTON,1950; The mean percent contribution of species KEMPF, 1972; WILSON, 1976), and may have of Pheidole in local ground ant communities, 700 or more valid species (E. O. WILSON,per - and the mean species richness of the Formici- sonal communication). Although it is at pre- nae and the assemblage as a whole (Fig. 1), re- sent impossible to place a name on the majority vea1 higher assemblage species richness from O of the taxa, especially in South America, spe- - 9 O N latitude, and a higher relative contribu- cies have clear morphological differences and tion of Pheidole in local assemblages from O - 9 can be fairly easily separated into morphospe- O S. When examined with respect to latitude cies. Complete worker dimorphism is a charac- without reference to hemisphere, the relative teristic of Pheidole, and behavioral specializa- contribution of species of Pheidole in local tions of the major caste may be inversely pro- ground ant assemblages declined significantly portional to their representation within the with increasing latitude (F = 26.255; P 0.001) worker population of the colony (WILSON, as did mean generic (F = 9.685, P 0.001) and 1984), although other researchers have not ob- mean species richness (F = 18.364; P 0.00 1 ) of tained this pattern (ARAB ef al., 1989; PATEL, the assemblages did (Table 1). These results 1 990; TSUJI,1990; LACHAUDefal., 1992). Spe- are in accord with previous studies (Kus~~zov, cies of Pheidole have diversified ecologies, in- cluding seed harvesters, general omnivores and 1957). predators, as well as mutualists in associations Likewise, the relative contribution of Phei- with many plants and Homoptera (KUSNEZOV,dole in local ground ant assemblages remained constant, irrespective of habitat examined (F = 1956; BRIAN, 1983; HOLLDOBLER& WILSON, 1990). Because of this, members of Pheidole 3.437; P 0.05), but not with respect to the type are found in nearly al1 types of terrestrial habi- of sampling method employed (F = 10.745; P 0.05) (Table 2). Species of Pheidole were signi- tats (WILSON, 1987). Recent efforts have been made to use spe- ficantly more abundant on baits than any other cies of Pheidole as an indicator group for sampling methodology. This was probably due to their efficient mass recruitment capabilities comparative diversity studies (BENSON & (WILSON, 1987), which allows them to domina- BRANDAO, 1987; MOUTINHO,1991). Here, 1 examine the relative contribution of Pheidole te fwd resources and effectively exclude other competitors. Both habitat and sampling met- species in 195 local ground ant assemblages of the Americas. hod influenced generic and species richness patterns of entire assemblages (Table 2). Soil and litter samples always had higher generic and species richnesses, while nest counts had DATA BASE the lowest (Table 2). Thus sampling method re- sulted in significant heterogenity of sampled Data are derived from a survey of literature, habitats for both species and generic richness and a complete listing, as well as criteria used (Table 2). Forests and closed habitats always for classifying major habitat types, can be had higher generic and species richnesses, whi- found elsewhere (FOWLER, in press). Because le grasslands and other open habitats had redu- ant assemblage size varied from a dozen to ced richnesses, resulting in significant hetero- more than 100 species per assemblage, 1 stand- genity of the sampled communities (Table 2). ardized local assemblages by taking the per- However, both habitat and sampling methods centage of Pheidole of the total assemblage spe- were not independent of latitude (FOWLER, in TABLE1. The latitudinal distribution of ground ant assembly species and generic richness, and the respective participation of species of Pheidole in local ant assemblages of the Americas. La distribución latitudinal de riqueza de especies y géneros en comunidades locales de hormigas de suelo, y la participación relativa de especies de Pheidole m comunidades locales de hormigas en las Américas.

Latitude % Pheidole Mean Genera Mean Species N- Richness

Mean Formicinae r ichness Latitude Mean % Pheidole North - Mean SR ) 40 richness 30-39 20-29 10-19 0-9 Equator - 0-9 1 0-19 1 20-29 -I 30-39 ) 40 + South - I I I I 1 111 #1*1 1 1 1 11 1)Il 0.4 0.3 0.2 0.1 O 10 100

FIGURE 1. The reiative contribution of species of Pheidole to local ground ant communities, and corresponding species richness of Formicine ants and the empire ant assemblage with respect to latitude. Data derived from 195 studies (FOWLERin press) La contribución relativa de especies de Pheidole en comunidades loceles de hormigas de suelo, y la wmspondiente riqyza de especies de hormigas Formicinae y de la comunidad total con respecto a la latitud. Los datos se han extraído de 195 estudios. 32 H. C. FOWLER ANNiS Dü BIOIOGfA, 19 (1993)

TABLE2. Mean relative representatim of Pheidole species in local ground ant communities of the Americas, and the respective mean number of genera and species in these communities with respect to the habitat sampled and the type of sampling method employed. La representación media relativa de especies de Pheidole en comunidades locales de hormigas de suelo en las Américas, y el número medio respectivo de géneros y especies en estas comunidades con relación al habitat de la muestra y el tipo de método de muestra empleado.

Habitat % Pheidole Total Genera Total Species

row crop 14.6 f 10.2 14.1 f 9.1 24.7 f 16.7 grass/open 12.5 f11.3 9.1 f 5.2 17.1 f 14.7 tree crop 23.5 f 14.5 12.1 f 9.5 21.6 f 22.0 forest/closed 14.7 f 14.5 16.4 f 11.3 36.8 f 34.7

F value 3.437 8.907* 7.781*

Type of sample

Litter/soil 13.9 f 11.3 18.2 f 15.2 38.4 f 41.7 Baits 22.1 f 16.0 12.2 f 8.7 31.1 f 31.5 Pit-fa11 trap 9.2 f 8.5 16.2 f 9.7 26.8 f 23.3 Nest counts 14.3 f 11.1 8.0 f 3.2 12.6 f 7.4 General 6.6 f 7.8 14.9 f 8.0 29.4 f 20.0

F value 10.745* 9.932* 12.664*

P c 0.05

press). In the Americas, nest counts are used ever, no significant relation was found between extensively in temperate areas, while baits are Pheidole relative assembly participation and la- used more intensively in the tropics (FOWLER, titude (Fig. 3). in press), probably due to the difficulty of loca- The Formicinae also presented a signifi- ting nests. Likewise, most studies in the tempe- cant inverse relationship with relative Phei- rate regions of the Americas have been conduc- dole species richness in local communities ted in open arid regions, while in the tropics (Fig. 4; Table 3). Furthermore, significant la- attention has focused upon crops and forests titude X Pheidole interactions were also (FOWLER,in press). found, with the later having a larger contri- The percent contribution of Formicinae in bution in local assemblage contributions: local ant assemblages demonstrated a positive % Formicinae = 0.267 (O Latitude) - 0.396 relation with increasing latitude (Fig. 2). How- (% Pheidole) -0.3 15 (O Latitude X % Pheido- AJiAIES DE BlOU)<;k19 (1993) REPRESENTATION OF PHEIDOLE IN AMERICA 33

TABLE 3. The relation of the percent taxonomic representation to the percent of Pheidole in local ground ant assemblages in the Americas. Regression model is % subfamily representation = intercept + slope (% Pheidole). La relación del porcentaje de representación taxonómica con el porcentaje de Pheidole en comunidades locales de hormigas de suelo en las Américas. El modelo de regresión es % de representación de la subfamilia = interceción + pendiente (% Pheidole).

Subfamily intercept Slope F value r2

Ponerinae 0.06 1* 0.171* 8.882* 0.046 Ecitoninae 0.016* -0.0 16 0.4 19 0.274 Pseudomyrmecinae 0.005 0.039* 4.626* 0.025 Dolichoderinae 0.086* 0.027 0.264 0.00 1 Formicinae 0.383* -0.859* 75.615* 0.290 Myrmicinae 0.457 0.562* 39.227* O. 176

FIGURE2. The relative contribution of Formicine ants in local ground ant assemblages of the 2 Americas as a function of latitude [% Formicine = -0.036 + 0.01 1 O latitude; r = 0.439; F = 144.677, P < 0.00011. Letters refer to the type of sampling method employed: N = nest countes; S = soil samples; L = litter samples; P = pit-fa11 samples; B = bait samples; and G = general collection techniques.

La contribución relativa de Formicinae en comunidades locales de hormigas de suelo en las Américas en función de la latitud [% Formicinae = -0.036 + 0.01 1 O latitud P = 0.439; F = 144.677, P c 0.00011. Las letras representan el tipo de método de muestra usado: N =censos de nidos; S = muestras de suelo; L = muestras de hojarasca; P = muestras de pit-faii; B = muestras con cebos; G = técnicas generales de recolección. 34 H. G. FOWLER ANWDE BH)LOC~A,19 (1993)

FIGURE3. The relative mtribution of species of Pheidole in local ground ant assemblages of the Americas as a function of latitude [% Pheidole = 30.585 - 0.481 " latitude; rZ= 0.095; F = 10.115, P = 0.0021. Letters refer to the type of habitat sampled: O = open; F = closed.

La contribución relativa de especies de Piddole en comunidades locales de hormigas de suelo en las Américas en función de ia latitud 1% PMáole = 30.585 - 0.481 latitud; r' = 0.095; F = 10.1 15. P = 0.0021. Las letras representan el tipo de hábitat de la muestra: O = abietto: F = cerrado.

FIGURE^. The relationship between the percentage representation of species of Formicine and Pheidole in ground ant assemblages of the Americas.

La relación entre el porcentaje de especies de Fonnicihae con Pheidole m comunidades locales de hormigas de suelo m las Américas. le) [r2=0.455, F = 26.202, P 0.0011 (Fig. 4). placement of the related myrmicine genus Relative species richness of the Pseudomyr- Myrmica as found in boreal Europe (VESPALAI- mecinae, Ponerinae and Myrmicmae increa- NEN & SAVOLAINEN, 1990). Interspecific territo- sed with increasing representation of Pheido- riality could reduce the number of ecologically le in local assemblages (Table 3), although distinct species to values smaller than the num- these effects were less than the inverse rela- ber of taxonomic species in the assemblage tionship found between Formicinae and Phei- (CODY, 1974; PONTIN,1982), and the use of dole. These results suggest that Pheidole con- higher taxonomic levels in this paper are justi- tributes proportionally more to tropical fied by the results. ground ant assemblages as species richness Biogeographic pattems are products of his- increases as evidenced by the positive slopes torical and ecological factors (ENDLER, 1982), of Pheidole contribution for the Pseudomyr- and an understanding of the causes and conse- mecinae, Ponerinae and Myrmicinae (Table quences of the relative contributions of each is 3), with a corresponding strong reduction of a major task of biogeographers. However, the- the Formicinae. se unexpected but robust results suggest that The negative relationship between Pheidole our knowledge of the nutritional ecology of and formicines suggests that these may be in- American ants is,at best, fragmentary and in- volved in interference or diffuse competition, complete. Furthermore, due to poor fossil reas found by VESPALAINEN& SAVOLAINEN (1990) cords of New World ants, it is at present impos- for other myrmice species in boreal European sible to try to separate histoncal from ecologi- ant communities. Another possible explanation cal factors. Nevertheless, the formicines are is that species richness patterns of both groups much more important in the higher North are linked through a stress gradient, with for- American latitudes, which may be due to their micines dominating more stressed environ- inferred Northem Hemisphere diversification ments, while Pheidole is more abundant in (KUSNEZOV,1953; BROWN, 1973). However, more favorable sections of the gradient. P. R. S. given the general dominance of Pheidole in MOUTINHO(personal communication) has ground ant communities of the Americas found an increasing faunal representation of (WILSON, 1987; BENSON & BRANDAO, 1987; Pheidole and a corresponding decline of formi- MOUTINHO, 1991 ), in species rich communi- cines in successional seres in the eastem Ama- ties, especially in the tropics, Pheidole may zon of Brazil, and suggests that this is due to regulate ant diversity (SAMWAYS,1990; AN- micro-climatic effects, thus reforcing the stress DERSEN, 1992), apparently by reducing formi- gradient hypothesis. However, only experimen- cine abundance. Although the foraging ecolo- tal studies can separate these conflicting hypot- gies and nesting and nutritional demands of heses. Pattems in contemporary ground ant as- most ant species are still unknown, these results suggest that species of Pheidole and for- semblages do not indicate Dolichoderinae dis- micines may be using the same resources, placement by Pheidole, as suggested by BROWN and open further questions which must be (1973) in evolutionary time. However, in eco- necessarily resolved in carefully designed logical time local exceptions may occur, espe- field experiments. cially with exotic species, such as Linepithema humile (Mayr) displacing Pheidole megacephala to an eventual spatial equilibnum in Bermu- da (LIEBERBURG et al., 1975), but these cases are exceptional in the Americas (FOWLER, ACKNOWLEDGMENTS 1992; FOWLERet al., in press). Because the effect of Pheidole participation in local comrnu- 1 gratefully acknowledge the comments of nities was even more intense than latitude on Edward O. Wilson, Sanford D. Porter, Law- the reduction of formicines, these results sug- rence E. Gilbert, Jonathan Majer, Jacques De- gest that in ground ant communities of the labie, Mike Karpaski, Marcelo Schlindwein, Americas species of Pheidole limit the formici- Alexandre Diniz Filho, Paulo Moutinho, nes (KUSNEZOV,1956), and not the inverse dis- Woodruf W. Benson, J.P. Lachaud, and an 36 H. G. FOWLER ANUDE BIOLOGh, 19 (1993)

anonymous reviewer on the manuscript. Sup- -, SCHLINDWEIN,M. N. & MEDEIROS, M. A. (in press): port was provided by the Conselho Nacional de Exotic ants and ant assembly simplification in Desenvolvimento Científico e Tecnologico of natural and agricultura1 areas of Brazil, with a Brazil (CNPq grants 3001 7 1/88-9 and review of the effects on native ant asserribla- 5006 14/9 1 - 1). 1 especially acknowledge the ges. In: WILLIAMS,D. F. (Ed.). Exotic ants: constant support and encouragement of Maria problems and perspectives. Westview Press, Thereza Fink, and Nozor Pinto for rendering Boulder. the figures. HOLLDOBLER, B. & WILSON,E. 0. 1990: The ants. Harvard University Press. Cambridge. KEMPF,W. W. 1972: Catálogo abreviado das formi- gas da Regiao Neotropical (Hym., Formicidae). Studia Entomol., 15: 3-344. REFERENCES Kus~~zov,N. 1953: Formas de vida especializadas y su desarrollo en diferentes partes del mundo. Du- senia. 4: 85-102. - 1956: A comparative study of ants in desert re- ANDERSEN, A. N. 1992: Regulation of "momentary" diversity by dominant species in exceptionally gions of central Asia and South America. Am. rich ant communities of the Australian seasonal Nat., 90: 349-360. tropics. Am. Nat., 140: 401 -420. - 1957: Number of species of ants in fauna of diffe- rent latitudes. Evolution, 1 1: 298-299. ARAB, A., LACHAUD, J. P. & FRESNEAU,D. 1989: Effets de la variabilité interindividuelle sur la taille des LACHAUD,J. P., PASSERA, L., GRIMAL,A., DETRAIN, A. répertoires comportementaux des deus sous- & BEUGNON, G. 1992: Lipid storage by major castes ouvieres de Pheidole pallidula (Formici- workers and starvation resistance in the ant dae, Myrmicinae). Actes Coll. Ins. Soc., 5: 251- Pheidole pallidula (Hymenoptera, Formicidae). 258. In: BILLEN, J. (Ed.) Biology and evolution of so- BENSON, W. W. & BRANDAO,C. R. F. 1987: Pheidole cial insects. 153- 160. Leuven University Press, diversity in the humid tropics: a survey from the Leuven. Serra dos Carajas, Pará, Brazil. In: EDER, J. & LIEBERBURG,I., KRANZ, P. M. & SEIP,A. 1975: Bermu- REMBOLD, H. (Eds.). Chemistry and biology of so- dian ants revisited: the status and interaction of cial insects. 593-595. Verlag J. Pepemy. Mun- Pheidole megacephala and Zridomyrmex humilis. chen. Ecology, 56: 473-478. BRIAN,M. V. 1983: Social insects: ecology and beha- MO~NHO,P. R. de S. 1991 : Composipio e diversida- vioural biology. Chapman and Hall. London. de de faunas locais de fomigas do genero Pheido- BROWN,W. L. Jr. 1973: A comparison of the hylean le Westwood (Hymenoptera: Formicidae) em and Congo-West Africa rain forest ant faunas. In: áreas florestadas. Dissertaciio de Mestrado. Uni- MEGGERS,B. J., AYENSU,E. S. & W. D. DUCK- versidade de Campinas, Campinas, Silo Paulo, WORTH, W. D. (Eds.). Tropical forest ecosystems Brazil. of Afiica and South America: a comparative re- PATEL,A. 1990: An unusually broad behavioral review. 161 - 185. Smithsonian Institution Press. pertory for a major worker in a dimorphic ant Washington, D.C. species: Pheidole morrisi (Hymenoptera: Formici- CODY, M. L. 1974: Competition and the structure of dae). Psyche. 97: 181-191. bird communities. Princeton University Press. PONTIN,A. J. 1982: Competition and coexistence of Princeton. species. Pitman Advanced Books. London. CREIGHTON,W. S. 1950: The ants of North America. SAMWAYS,M. J. 1990: Species temporal variability: Bull. Mus. Comp. Zool., 104: 1-585. epigaeic ant assemblages and management for ENDLER,J. A. 1982: Problems in distinguishing histo- abundance and scarcity. Oecologia, 84: 482- rical from ecological factors in biogeography. 490. Am. Zool., 22: 441-452. Tsup, K. 1990: Nutrient storage in the major workers FOWLER, H. G. 1992: Native faunal simplification by of Pheidole ryukyuensis (Hymenoptera: Formici- an exotic ant following hydroelectric dam cons- dae). Appl. Ent. Zool., 25: 283-287. truction in northeastem Brazil. Ciencia e Cultura, VESPALAINEN,K. & SAVOLAINEN,R. 1990: The effect 44: 345-346. of interference competition by formicine ants on - (in press): Ground ant (Hymenoptera: Formici- the foraging of Myrmica. J. Anim. Ecol., 59: dae) assemblages of the Americas. 1. Biogeogr. 643-654. WILSON,E. 0. 1976: Which are the most preva- noptera: Formicidae). Behav. Ecol. Sociobiol., lent ant genera? Studia Entomol., 19: 187- 16: 89-98. 200. - 1987: Causes of ecological success: the case of the - 1984: The relation between caste ratios and divi- ants (The Sixth Tansley Lecture). 1. Anim. Ecol., sion of labor in the ant genus Pheidole (Hyme- 56: 1-9.