MUSEUM COLLECTIONS ARE BETTER THAN EXPECTED: MINIMAL EFFECTS OF COLLECTION BIAS IN DETECTING INSECT HERBIVORY ON FOSSILIZED LEAVES

Museum collections represent a wealth of paleontological data, and yet are almost certainly subject to biases that must be accounted for if they are to be used in studies of paleoecology. The emphasis of many museum collections is on exceptional and rare specimens, as opposed to those that represent ecological or taphonomic completeness. This "high-grading" results in museum collections that likely misrepresent the actual abundance or richness of taxa in an assemblage. Additionally, leaves that have damage made by herbivorous insects may be discarded by collectors who might categorize them as “damaged”. To examine the role of collection bias in fossils floras, museum specimens from the middle Miocene Clarkia Formation (15 mya) were compared to newly collected field samples from the same locality.

Fossil leaves from the Clarkia Flora are well known for their excellent preservation and have been collected since the 1970’s. For this study, 158 fossil leaves from the University of Idaho’s Clarkia collection were examined. Taxonomic and herbivory data were collected from leaves from the P37 locality. In addition to museum specimens, new collections were made from two sites at this locality. Field samples consisting of 200 potentially identifiable leaves or leaf fragments were obtained. The number of leaves per sample with insect damage and the number of leaves with five basic damage types (holes, margin feeds, skeletonizing, leaf mines and galls) were recorded. Overall, the field samples had 18 of 33 (55%) leaf taxa in common with the museum collection. In addition, the field samples were not significantly different from the museum collection in terms of the prevalence of insect damage (R²=0.03, P=0.43), nor in the incidence of leaf mines (R²=0.00, P=0.85) and galls (R²=0.00, P=0.99). Therefore, we found the previously collected and potentially “high-graded” material to be equivalent to the newly collected material. This shows that for the Clarkia flora, samples housed at the University of Idaho do not have detrimental amounts of collection bias and they can be used in studies of herbivory that require more ecologically complete samples. This also suggests that other museum collections may not be as biased as previously thought and that many more floras may be suitable for studies requiring relatively unbiased samples.