Forensic entomology is characterized by the use of insects, mites and other arthropods in legal procedures. Entomologic information allows the post morten interval (PMI) estimates to be obtained with reasonable accuracy, even after months of death. The study of cadaveric succession includes corpse necrophagous fauna analysis, which is based on the levels of decomposition. The fauna may vary because each period offers ideal conditions for the development of certain species. The development of the eggs and larvae is extremely predictable, and detailed observations of the biological stages of several insect species as well as their succession in the decaying corpse may reveal important information regarding the place and time of death [1] .

Several traps can be used in forensic entomology to analyze the cadaveric entomofauna and to determine the relationship of the entomological succession with the phases of decomposition [2] -[5] . These traps are primarily characterized by the assembly of a rectangular iron framework that is usually clothed laterally by organza with an opening at the bottom to enable insects to enter the trap. Insect collection is generally performed manually using an entomological net wrapped over the top of the trap. However, some experiments do not use traps, and the insects are collected at the cadaver using insect hand nets [6] -[10] . For both methods, the researcher can collect the insects under the organza in addition to collecting the flying adults at the carcass. This practice may be harmful to the researcher, because these insects are known to be pathogen vectors [11] [12] .

A similar trap was described by Shannon (1939) [13] , and it was used primarily for the capture of day-flying sylvan mosquitoes and insects attracted by any type of bait. This trap consists of a rectangular compartment that is covered by muslin and four sides composed of mosquito netting. The trap is approximately 65 cm above ground with a large space below for the insects. The mosquitoes are caught with a suction tube inserted into a small opening in the side of the trap.

These traps remain assembled for undetermined lengths of time, thus allowing the collection of insects to visit the cadaver (or bait) throughout the day. However, without a bottle collector, the number of captured insects is low because many insects eventually die and fall to the ground and the nocturnal insects are not captured. In many experiments, oil is used as a fixative substance, which may alter the colors and structures of the insects that are prepared for identification.

In this study, we describe and test a modified Shannon trap for forensic entomology experiments. This trap has the following improvements: 1) collects the visiting insects during the entire decomposition process without prejudice based on the quantity and quality of the collected specimens, 2) prevents direct contact with the collected insects and 3) optimizes the time because of easy implantation that enables collection within a short period.

The trap proposed in this study consists of a voile textile tent fastened to the nozzle of a flask and fixed to the ground by four stakes, forming a pyramid with sides 10 to 30 cm from the ground to enable insect entry. The flask is fastened to two fixed points by its upper part, and there is an opening in the middle to which a collecting vial containing 70% gel alcohol is tied. Therefore, the collection is performed externally to the trap, and the insects are found adhered to the alcohol. The instructions for assembly are shown in Figure 1.

The trap was tested in experiments conducted in two distinct regions in Brazil between January and August, during the summer and winter seasons, respectively. The experiments were performed in the cities of São Paulo (at the campus of the Faculty of Medicine at the University of São Paulo, location 23˚33'S and 46˚40'W) and Peruíbe (in the EEJI—Estação Ecológica Juréia-Itatins, location 24˚22'S and 47˚01'W).

Four experiments were conducted, two during each season. The duration of each experiment was standardized to 20 days despite the variations in the decomposition periods of the carcasses at the two locations.

Swine carcasses weighing approximately 12 kg were placed directly on the ground inside 70 × 50 × 50 cm cages in shallow pits approximately 10 cm deep. This procedure was performed to protect the carcasses from vertebrate predators and to allow for adequate colonization. The cages were closed and strung in the surrounding vegetation to avoid predation as previously discussed. The trap was assembled above the cage to verify its efficacy and applicability for both experimental and forensics use. The collections were performed on alternating days at both locations.

After collection, the samples were transported to the laboratory and later grouped according to the decomposition phase. With the objective of evaluating the practical applicability of the trap in routine forensics, two collections with a one-hour interval were performed (the first collection at noon and the second collection at 1 p.m.). After this procedure, the collection vial was repositioned in the trap for additional collections.

A total of 10,833 arthropods that belonged to 53 families of insects were collected. The majority of the insects were from the order Diptera (96.9%), but specimens from the orders Coleoptera (2.12%), Hymenoptera (0.71%), Lepidoptera (0.14%), Hemiptera (0.10%) and Orthoptera (0.03%) were also observed. Table 1 shows the number of insects collected at each decomposition stage, and a quantitative analysis of the collected insects according to their taxonomic level of family is presented in Table 2.

The modified trap (Figure 2) retained the insects, and it was resistant to inclement weather. All of the collections were successful.

To test the functionality of the trap for experimental studies, the trap was maintained above the carcass during all phases of cadaveric decomposition. The collected insects were directly related to the phase during which they were found, thereby allowing analysis of the pattern of entomological succession. In addition, quantification

of the collected insects made it possible to identify the abundance of the orders and families.

An evaluation of the efficacy of the number of specimens captured was performed by comparison with results from other studies that used similar traps or insect hand nets as shown in Table 3 [2] -[10] . Based on this analysis, our trap was more efficacious because of the increased number of insects captured during the process of decomposition and the variety of species captured.

To test the applicability of this trap for use in forensics, the quickness of assembly and the potential to collect insects in a short period were assessed. The results of both evaluations were positive because the implantation of the trap was simple and an estimated 30 to 80 insects were collected per hour.

^*EEJI-S—Peruíbe experiment during the summer. SP-S—São Paulo experiment during the summer. EEJI-W— Peruíbe experiment during the winter. SP-W—São Paulo experiment during the winter.

*EEJI-S—Peruíbe experiment during the summer. SP-S—São Paulo experiment during the summer. EEJI-W— Peruíbe experiment during the winter. SP-W—São Paulo experiment during the winter.

We conclude that the trap showed favorable results. Its low cost, ease of manufacturing and efficient collection and preservation of the specimens captured facilitates further identification and characterization of necrophagous fauna. Furthermore, this study demonstrated that the trap met the expectations of feasibility related to the duration of the experiments, the number and quality of the specimens captured and the richness of the species collected. Therefore, this trap can serve as a complementary resource for forensics practice and experimentation.

We thank Prof. Dr. Silvio Nihei for help reviewing the text, and Julia Albuquerque and Igor Caricaturas for help with the language and figures.