Supplementary notes on the Gentry 0.1 ha transect dataset

 

Anyone using these data should be aware of the following caveats, ambiguities, and sources of error:

 

1) The original data were recorded in an unusual “species (stems)” format. 

 

For each subplot (=2 m x 50 m line), each species was noted only once, and the number of individuals tallied, along with the diameters of all stems > 2.5 cm dbh.  Because an individual may have more than one stem, this system results in an ambiguous mapping of stems onto individuals: there is no way to know which stem or stems go with which individual. Thus, any calculation involving both stems and individuals are invalid for this dataset (for example, you cannot calculate number of stems per individual).

 

2) How to count individuals.

 

When downloaded from the SALVIAS database, these data are presented as a one-line-per-stem flat file.  The field OBSERVATION_ID identifies a single observation of a species in a single subplot (50 m x 2 m line).  All stem measurements (stem_dbh) bearing the same value of OBSERVATION_ID pertain to that particular species-subplot observation.  The number of individuals observed for a given species in a given subplot  (“number_of_individuals”) is recorded on the same line as the first stem observation for that species in that subplot.  Number_of_individuals for any additional stems is set = 0.  This permits individuals to be totalled for plots, subplots, or taxon simply by grouping on the appropriate field and summing number_of_individuals.

 

3) In many cases, multiple vouchers specimens were recorded for a given species-subplot observation.

 

For many individual records in the Gentry dataset bearing more than one specimen voucher, it may be impossible to update species determinations from specimens.  There is simply no way to know which is the “correct” specimen in cases where determinations differ among vouchers for the same record. 

 

4) Many collection numbers are in error.

 

A high percentage of collection numbers of voucher specimens in this dataset seem to be in error.  Upon checking these numbers in the TROPICOS database, you will notice that many refer to obviously incorrect taxa (e.g., herbaceous aquatics) and/or are from obviously wrong locations (in some cases, from countries other than where the transects was collected).