24 



SHELVE DISTRICT 



The number of individuals attributable to any species represented in the collection 

 is small enough to warrant not only the invariable use of small sample techniques but, 

 for many taxa, nothing more than the mean (or mode) and range of variability. 

 Continuous univariate characters which are assumed to be normally distributed, 

 like the distances at which folds originate anterior of umbones or the wavelengths of 

 ribs at a given distance antero-medially of the umbo, have been compared by the 

 Mest. Discontinuous univariate characters, like the number of costae ornamenting 

 valves or the relative branching of costellae enumerated according to Bancroft's 

 notation (Bancroft 1945 : 186 ; Williams 1962 : 77), have been compared by x 2 tests, 

 or contingency or two-by-two tables dependent on the size of samples. 



Continuous variables derived by measuring the shell or its several parts, like the 

 cardinalia, dental plates or muscle impressions, are expressions of incremental 

 proportional growth, which may or may not be allometric, and all show very high 

 positive (or rarely negative) correlations with one another. The significance of any 

 differences in estimates of such parameters has been determined by statistical analysis 

 of a series of bivariate distributions. The method is adequate for the comparison 

 of such closely inter-related estimates of shell growth. It also has the advantage of 

 not obscuring the identity of those characters that underwent changes in the rate of 

 growth during cladogenetic or phylogenetic speciation. The development of sig- 

 nificantly longer dental plates in one of a number of compared species, for example, 

 will be apparent whether the lengths of plates are consistently paired with the 

 lengths, maximum widths or depths of the pedicle valves in compared samples. 



In estimating variation in outline, shape and relative size of the brachiopod 

 skeleton, very many measurements in almost any plane or direction may be taken, 

 provided the same disposition is adopted in measuring all individuals in compared 

 samples. However, certain orientations for the measurement of shells or their 

 impressions have become widely favoured by brachiopod researchers over the years 



Fig. 3. Diagrammatic representation of moulds (and adherent shell) of Orthambonites to 

 show the orientation adopted for measuring certain dimensions including : length (1), 

 maximum width (w) and thickness (th) of a valve, the wavelength of a costa (wv), the 

 length of dental plates (dl), the length (Isc) and width (wsc) of muscle scars, the length of 

 a median ridge (Is), and the length (lc) and width (wc) of brachiophore bases. 



