Size of data worksheets

There are no fixed size limits for arrays within PRIMER 7, simply an overall limit determined by the amount of available real memory on the computer. There will, of course, be significant time constraints for some of the more compute-intensive routines, and there is a limit to how many samples it is sensible to try and view at the same time, in ordination plots for example, but it is a viable strategy to place all related data into a single workspace, and data of the same type (e.g. counts for a specific faunal assemblage in a complete set of samples) into a single worksheet. Defining a structure of factors on the samples (see Section 2) will allow the selection of subsets from that worksheet, or averaging over replicates (or factor levels), needed for a specific analysis. Continually improving computational power makes it possible at least to hold and manipulate arrays with thousands of species (typically OTU’s in microbial applications) and thousands of samples, even if successful analyses will often involve targeted selections or averages of samples. Whether matrices are input as samples (rows) $\times$ variables (columns) or variables $\times$ samples is not of relevance: PRIMER simply needs to be told whether the samples are the rows or the columns. Note that transposing a worksheet within PRIMER is possible (by Tools>Transpose), but will not correct a mistaken attribution on entry – if the rows have been incorrectly called ‘samples’ during the Open dialog then, after transposing the worksheet, the columns will still incorrectly be called ‘samples’. Instead, the mistake is corrected by Edit>Properties and taking Samples as•Columns.