Estimating populations of Idiarthron subquadratum (Orthoptera: Tettigoniidae) using mark-recapture methods in coffee plantations in Chiapas, Mexico

Idiarthron subquadratum (Orthoptera: Tettigoniidae) is a sporadic pest of coffee in parts of Mesoamerica. Mark-recapture techniques were employed to determine the adult population size in coffee plantations in the municipality of Siltepec, Chiapas, Mexico, where the insect has caused crop losses of around 50% over the past decade. Eight experimental blocks 50 by 50 m were defined in an area of 2,500 m(2). A trap was designed consisting of a bamboo tube 30 cm in length and closed at one end. Trials indicated that insect trapping efficiency was not improved by the presence of baits. Traps placed above the ground in coffee bushes or as banana leaves placed on the side of an experimental gauze cage captured significantly more L subquadratum individuals than traps placed on the ground. Tests with different types of paint combinations for marking insects revealed that nitrocellulose paint was durable, water-resistant and nontoxic to L subquadratum. Mark-recapture experiments were performed at 48-h intervals over a 1-mo period. Concurrently, direct nocturnal observations of I. subquadratum individuals feeding on coffee leaves and berries were made on four occasions during the study period. The prevalence of I. subquadratum feeding damage was also quantified at three sample dates during the study. Jackson's negative and positive index, Fisher-Ford model, Lincoln-Petersen index, and Jolly-Seber's stochastic model gave statistically similar mean population estimates. All of these mark-recapture models were also statistically similar to estimates generated by direct nocturnal counts of insects observed feeding on coffee bushes. In contrast, Bailey's triple capture model gave a population estimate significantly lower than the other models. L subquadratum feeding damage to coffee berries was significantly greater in blocks interplanted with banana. We conclude that mark-recapture methods combined with analysis using the Lincoln-Petersen model are simpler and less time consuming than direct nocturnal observations and give quantitatively similar population estimates.