Richness and composition of gall-inducing arthropods at Coiba National Park , Panama

Interest in studying galls and their arthropods inducers has been growing rapidly in the last two decades. However, the Neotropical region is probably the least studied region for gall-inducing arthropods. A study of the richness and composition of gall-inducing arthropods was carried out at Coiba National Park in the Republic of Panama. Field data come from samples obtained between August 1997 and September 1999, with three (two-week long) more intensive samplings. Seventeen sites, representing the main land habitats of Coiba National Park were surveyed. 4942 galls of 50 insect and 9 mite species inducing galls on 50 vascular plants from 30 botanical families were colleted. 62.7% of the galls were induced by gall midges (Diptera, Cecidomyiidae), 15.3% by mites, Eriophyidae, 8.5% by Homoptera, Psyllidae, 6.8% by Coccidae and 5.1% by Phlaeothripidae (Tysanoptera). The host plant families with the most galls were Myrtaceae with seven, Bignoniaceae with five and Euphorbiaceae, Fabaceae and Melastomataceae with four. Leaf galls accounted for about 93% of collected galls. Most leaf galls were pit/blister galls followed by covering and pouch galls. Gall richness per collecting site was between 1 and 19 species. Coiba’s gall diversity is discussed in relation to data available from other tropical sites from continental Panama and the Neotropical region. Our results support the idea that it may be premature to conclude that species richness of gall inducers declines near the equator. Rev. Biol. Trop. 56 (3): 1269-1286. Epub 2008 September 30.

Many groups of insects and mites, estimated in a range from 21 000 to 211 000 species, with an average of 132 930 species (Espírito-Santo and Fernandes 2007), are capable of inducing galls on plants, which are structures that involve active differentiation and growth of plant tissues providing shelter, nutrition and protection for the gall-inducing organism (Meyer 1987, Nieves-Aldrey 1998, Stone and Schönrogge 2003).
Galls and gall-inducing arthropods have been the object of research for the last three centuries by dozens of naturalists and scientists worldwide (Mani 1992, see Raman et al. 2005 for a recent synthesis).Earlier information was compiled in a classic work by Houard (1908Houard ( -1913Houard ( , 1922Houard ( , 1933Houard ( , 1940)).The most important groups of gall-inducing insects are Diptera (Cecidomyiidae), Homoptera (Cicadoidea, Psylloidea, Aphidoidea and Coccoidea), Thysanoptera and the hymenopteran families Tenthredinidae and Cynipidae.Mites inducing galls belong mostly to the family Eriophyidae.These galling-insect groups have different distribution patterns, and their distribution may be correlated with the breaking up and movement of continents (Gagné 1984).Geographic and climatic factors could also have determined the distribution patterns of gall-inducing arthropods (Schlinger 1974).The most important family of gall insects is Cecidomyiidae (Diptera), predominant in almost all zoogeographic regions, both tropical and temperate.Cynipidae and Tenthredinidae are mainly holarctic, gall-inducing thrips (Thysanoptera) are restricted mostly to Western Asia and gall-inducing psyllids and coccids are predominant in tropical regions (Felt 1940, Gagné 1984).
Interest in studying galls and their arthropod inducers has been growing rapidly in the last two decades (Raman et al. 2005).However, the Neotropical region continuous to be a poorly studied region for gall-inducing arthropods in general (Espírito-Santo and Fernandes 2007).The most important references are the classic work by Houard (1933), the book by Gagné (1994) on gall midges (Cecidomyiidae) and local studies in Brazil (Maia 2001, 2005, Gonçalves-Alvin and Fernandes 2001, Maia and Fernandes 2004, Costa De Oliveira and Maia 2005, Urso-Guiamares and Scareli-Santos 2006); Costa Rica (Hanson and Goméz-Laurito 2005); Mexico (Cuevas-Reyes et al. 2004a, 2004b) and Panama (Medianero et al 2001).Future research in the Neotropical region will likely change the biogeographical patterns assumed until now for gall inducing insects (Espírito-Santo and Fernandes 2007).
The aim of this study is to estimate the richness and composition of gall inducing arthropod species in an unexplored insular tropical habitat off the Pacific coast of Panama, and compare them with data available from continental Panama.

Study area:
The research was performed at Coiba National Park (World Heritage Site), as a part of the Joint Hispanic-Panamanian Program for National Parks.Coiba National Park is located in the Panamanian Pacific Ocean (7º39´-7º18´ N & 81º53´-81º35´W), 22 Km from the mainland, and encompasses an area of 270 125 ha, of which 53 528 ha are insular and 216 543 are marine.The insular area is composed of Coiba (50 314 ha), the largest island in tropical Pacific America, eight other minor islands and 30 islets.The maximum altitude in the Park is 420 m on Coiba Island.Annual rainfall is 3 483 mm.The yearly average humidity is 80% and the mean temperature is 25 ºC.The area is characterized by an intense dry season, from December to April and a rainy season from May to November.The natural vegetation is well conserved, with about 80% forest coverage.An untouched tropical moist and wet forest occupies 60% of the park (Castroviejo 1997).The island of Coiba became separate from the mainland about 12 000 to 18 000 years ago and houses many endemic species of animals and plants.The rough forested terrain has served to make the island inhospitable and preserved, but even more so the presence of a prison, consisting of some twenty camps scattered along the coastline, that was in operation from 1918 to 2004 (Castroviejo 1997, Fontal andNieves-Aldrey 2004).
Field and laboratory work: Field data were obtained from several samples between August 1997 and September 1999 with three more intensive samples (two-week long) in August 1997, July 1998 and August 1999.The sampling method consisted of transects during not less than one hour (see Price et al. 1998).Along the selected collecting sites all plants were searched for galls.Seventeen sites representing the main terrestrial habitats in Coiba National Park were surveyed (Table 1, Fig. 1).Plants with galls were collected and photographed.Specimens of galled plants were dried and put into a herbarium collection, where plants were identified later.Samples of the galls were also stored in 70% ethanol to allow further dissection and identification.Data of altitude and geographic position was taken in the field (Table 1).
The galls were dissected in the laboratory for adult and larval identification to family level.Gall-inducing species were separated by the external morphology of the galls, the larvae and the host plant.It is assumed in a study like this that gall morphotype is unique for each gall-inducing species (Mani 1964, Ananthakrishnan 1984).Yet a very conservative methodology was employed (see Hanson and Goméz-Laurito 2005); galls found in petioles, in leaf blades, as well as in the nerve of a given leaf, and those without differentiation in plant species of the same genus were considered as of the same gall-inducing species.
Data analysis: Gall-inducing arthropod species richness (S) was calculated based on the number of different galls found and the alpha diversity index (α) was estimated (Fisher 1943): S is the number of species in a sample, N is the number of individuals in a sample and α is the diversity index.This index is not influenced by the two limitations of other known indeces because it is independent of sample size and it does not give more weight to species abundance (Wolda 1983).
To verify sampling efficiency we constructed species accumulation curves with the estimator Chao 1, Bootstrap and Jacknife 1 (100 random).The species accumulation curve is based on individual-based assessment protocol (see Gotelli and Colwell 2001).We used the software EstimatesS® 8 (Colwell 2006).

Taxonomic and faunistic considerations:
A total 4 942 galls of 59 species, 50 insects and 9 mites, associated with 50 species of plants from 37 genera and 30 families were collected at Coiba National Park.It was not possible to identify the genus of nine host plant species  2).The plant families hosting the most gallinducing species were Myrtaceae with seven, Bignoniaceae with five and Euphorbiaceae, Fabaceae and Melastomataceae with four (Table 3).The plant species with the greatest diversity of galls was Calophyllum longifolium (Clusiaceae) with three, seven species were found with two galls (Amphitecna latifolia, Mansoa sp., Acalypha diversifolia, Casearia commersoniana, Lacistema aggregatum, Cassipourea elliptica and Pouteria foveolata) whilst only one gall-inducing species was found infesting two species of Ficus.Alpha diversity was 9.4 but, species accumulation curves did not become asymptotic suggesting than there are more gallinducing species at Coiba National Park than those collected in this study.Gall-inducing species number increases with individual sample (Fig. 2).Gall richness per collecting site was between 1 and 19 species.71% of all species were found at only one site, 22% in two, 5% in three, 2% in four, while no species was present in more than four sites.The gall inducing arthropod community studied shows a high complementarity among collecting sites.
Our results show than the moist and wet tropical forests of Coiba National Park revealed a rich diversity of gall-inducing arthropods, with species richness numbers higher than those registered in continental Panama (see Price et al. 1998) and other parts of the Neotropical region (see Cuevas et al. 2004a, Costa De Oliveira andMaia 2005) (Table 4).Although we did not collect all species in Coiba Island, the total of fifty nine species, with a range of one to nineteen gall inducer species per site, in Coiba National Park is more that the twelve proposed for mesic vegetation in the literature (see Price et al. 1998).Coiba National Park diversity is similar DISCUSSION Gall-inducing arthropods have been more thoroughly sampled in temperate regions than in tropical regions.Indeed, from the perspective of biodiversity inventories, we have barely scratched the surface of the tropical faunas (Hanson and Goméz-Laurito 2005).A global pattern in local number of gall-inducing insects was described recently (see Price et al. 1998).
The study pointed to a greater richness of gallinducing insects in warm dry regions, while temperate and tropical areas with humid, mesic vegetation showed lower gall richness indices.However, as recently acknowledged by Espírito-Santo and Fernandes (2007), these results could be affected by an under-sampling of tropical rain  to that found at San Lorenzo (α = 9.3), a six ha rain forest area on the Panamanian mainland (Medianero et al. 2003).Our results suggest that, according to insular biogeography theory (MacArthur and Wilson 1967), that Panama's continental territory, where close to 11000 species of plants have been identified (Correa 2001), as well as other sites in the Neotropical region, must be richer in arthropods than indicated by recorded data in literature.Certain climatic and edaphic factors can produce higher than expected gall-inducing species richness in certain warm temperature regions.Plant diversity is the predominant factor producing high gall-inducing species richness in wet tropical regions (Yukawa et al. 2001;Hanson & Gómez-Laurito 2005).Our results support the idea that it may be premature to conclude that species richness of gall inducers declines near the equator (Hanson and Gómez-Laurito 2005).Furthermore there is an urgent need for more investigation of gall-inducing arthropods in the Neotropics, a region where as many as 20.000 gall-midge species (Cecidomyiidae), could exist (Espírito-Santo and Fernandes 2007).

ACKNOWLEDGMENTS
This work was supported by the Spanish Agency of International Cooperation (AECI) and also partially funded by the research project DGES PB97-1241, of the Spanish Ministry of Education and Culture to JLNA.Thanks to the Panamanian Authority of the Environment (ANAM) for collecting permits.We thank all the staff of AECI in Panama and ANAM in the Biological Research Station of Coiba, as well as many colleagues from The Natural Sciences Museum and the Royal Botanical Garden in Madrid, his friendly help and support in several ways who made possible this research.Special thanks to Dr. Santiago Castroviejo who was the direct responsible of conceiving and initiating the Coiba Inventory Research Project.

Fig. 1 .
Fig. 1.Map of Coiba National Park, showing sampling sites.The four groupings correspond to the four different plant communities at Coiba Island.
Comparison of gall inducing arthropods species richness of Coiba Nacional Park and literature data from the Neotropical Region

TABLE 1
List of sampling sites data at Coiba Nacional Park mostly Myrtaceae.62.7% of the galls were induced by species of the family Cecidomyiidae, 15.3% by Eriophyidae, 8.5% by Psyllidae, 6.8% by Coccidae, 5.1% by Phlaeothripidae; one species corresponding to 1.7% could not be determined.93.1% of the galls were found on leaves, 3.4% on stems and 5.2% on buds (Table