The role of voucher specimens in validating faunistic and ecological research

Lack of vouchers renders published results unverifiable

McCorquodale (2001) recorded several new Ontario records of Cerambycidae based on old vouchers (see previous section on long term studies) but also identified a number of questionable published records for which no museum specimens could be found. As a result, there was no way to confirm the past occurrence of these species in Ontario.

In contrast to McCorquodale’s (2001) convincing illustration of the value of vouchers, another paper in the same issue of the same journal (Paquin and Dupérré 2001) recorded many new North American, Canadian and Quebec records in the course of a large biodiversity survey of boreal forest Coleoptera. Unfortunately, there was no indication as to where (or even if) voucher specimens were deposited and, as a result, there is no way for other workers to confirm the identity of those species in future.

In addition to documenting several misidentified species in studies that did match their sequence data to known voucher specimens (see above), Ruedas et al. (2000) found that 73% (41 of 56) of the papers they surveyed in Molecular Phylogenetics and Evolution did not link their sequence data to a voucher specimen at all; thus there is no way of knowing whether the species identifications were correct or not.

Although some errors in past biological control programs can be traced and corrected, as discussed above, many early biocontrol introductions were not usually accompanied by deposition of voucher specimens (Sabrosky 1955, Danks 1988, Huber 1998) so there is no way to know what went wrong.

The Canadian Nature Federation’s lady beetle survey in the 1990s (www.cnf.ca/beetle/index.html) was, unfortunately, a failure from a scientific perspective because the nationwide register of coccinellid species was based almost entirely on sight records submitted by non-specialists. Participants were not encouraged to collect specimens to serve as vouchers and the survey was rife with apparent misidentifications of species that cannot be corrected by subsequent examination of specimens (Marshall 2000). Thus, the data from the survey are useless for rigorous scientific analysis. In contrast, a more recent amateur initiative in eastern Canada, the Atlantic Dragonfly Inventory Project coordinated by P.M. Brunelle, has the potential to contribute to considerable scientific research because over 90% of the records in the species database are supported by voucher specimens.

“Recommendations” in the editorial policy of refereed journals on deposition of vouchers (Table 1) obviously have little impact on most authors. In three randomly selected issues of The Canadian Entomologist published in 2002, 30 papers involved species-level identification of insects, but only nine (six of which were taxonomic papers) mentioned deposition of voucher specimens. In two randomly selected issues of the Annals of the Entomological Society of America for 2002, 31 papers dealt with named species but only nine (eight of which were taxonomic) mentioned voucher deposition. The numbers are similarly discouraging for other journals.

Most identification work performed by freelance consultants or contractors is not documented by deposition of vouchers. This is particularly disturbing because many consultants are not trained specialists in arthropod identification and, in the absence of vouchers, their work cannot subsequently be verified by specialists. In many cases, such identifications subsequently checked by specialists have been found to have very high error rates (Danks and Winchester 2000, Marshall 2000). Given that the specimen identifications derived from contract work often are used in making decisions on environmental impact assessment, land use planning and conservation priorities, the potential negative implications of not having vouchers available for verification are significant and far-reaching.


Page updated on Feb 23, 2014