The key on this project is of the 'multiple entry' type using the IdentifyIt software. It is an identification system based on a matrix describing each species as a series of character states. Identification is done by entering a 'search pattern' that describes the specimen as a combination of character states. The software then compares the search pattern to the species in the file and gives the percentage match for each species as well as 'the next best character' to add to the search pattern in order to complete the identification. That means, the users can freely select among the available characters for the identification of the specimens at hand. Each character allows the decision for one or more character states. Theoretically, the keys allow the identification of a species based on only a few or even one character state. However, to achieve a reliable identification and avoid confusions by superficial similarities, certain decisions for some major characters have to be made. How to find these characters is explained in one of the next sections (How to start an identification).

Apart from identification the data files might also be useful for viewing species descriptions and comparisons (for this purpose, press the buttons "examine" and "compare").

This and the following sections explain some of the problems concerning the present key and the identification of Agromyzidae in general. However, it is not a help file explaining the software features. A general guide to the IdentifyIt software can be found in the help menu.

With altogether 111 species this key covers only a small fraction of about 2,750 agromyzid species (Henshaw and von Tschirnhaus, in prep.) worldwide. Many species, which were not included here, might occasionally occur on cultivated plants or could do so in the future. However, of many agromyzids the biology and host plants are unknown. Therefore, the users of the key must be aware of the possibility to have a species at hand that is not incorporated in the key. Therefore, after each identification a meticulous comparison of the specimens with the species description is necessary to avoid misidentification.

Agromyzidae are known as taxonomically difficult. Certain species can only be recognized by tiny differences of the male genitalia. Often identification requires only the comparison of drawings or pictures. Not all species can be described by characters that are easily included into a key.

In many cases only an approximation to the correct species can be achieved by the IdentifyIt tool. The set of potential candidates must be carefully compared using the descriptions and illustrations provided in the species module. In any case of uncertainty further literature or specialists should be consulted.

The data on some very similar species were grouped together into separated identification files (Agromyza-nigripes-group, Japanagromyza, Tropicomyia and Napomyza). These subgroups can be identified in the main identification file (Agromyzidae ID). Subsequently, detail files can be opened either separately or from inside the main identification file.

Some critical characteristics of the present keys and software

"Unknown" is equivalent to a defined character state

If in one of two taxa the state of a character is 'unknown', IdentifyIt treats the information as if both taxa have different known states of the same character. This can bias the resulting identification. Since, in the case of missing information it is unknown if the taxa are identical or not. Using the key one must be aware that the tolerance for unknown information is insufficient.

Equal weight to each character

The matrix key allows the identification of species based on a small selection of the available characters. Yet from conventional taxonomic experience ensues that some characters are more reliable than others are. This can be due to variability among the species, the indistinction of the character states, preservation methods and artifacts or different levels of information content of characters. However, in the matrix keys of LINNAEUS II all characters are coded equally. Thus, the users of the key should use the module critically.

Generally the number and position of microtrichia (bristles, setulae), color patterns, body surface should be handled with care. In cases of doubtful decisions the character should be skipped and further characters should be studied. If several cases of indistinct character states in some species two or more mutual conflicting characters were coded.

The male genitalia should be prepared and considered for identification whenever possible. Their complex structure often prevents strong misidentifications. Some species groups show hardly any difference except in the genitalia.

Limited text opportunity

For character descriptions the software allows only a limited number of characters (letters) per taxonomic character and state. The limited space made the use of abbreviations and sometimes too brief descriptions inevitable. For this reasons an annotated list of the characters used in the keys is given on the next pages (Main key - Character list and Detail keys - Character list).

Illustrations of characters or character states

As far as possible, terms relevant for morphological structures are explained by pictures in the key. Remaining difficulties can be solved by the glossary or a full-text-search.

Missing taxa due to the lack of information

The list of taxa treated on the CD-ROM but excluded from the identification files can be found in one of the next sections (abbreviations and missing taxa).

To identify the species missing in the keys, check the species cards for similar and related species above and compare the descriptions and images of similar species with each other; check the host plant module for alternative species on the same host plant.