8 Journal of the F.M.S. Museums. [Vol. VII, 



(2) There is probably a small error due to blacklash or 



spting in the dividers; this error is positive and 

 is relatively greater in the smaller than in the 

 larger measurements. 



(3) In the case of the measurement of total length old 



skulls may give a longer measurement than that 

 really representing their morphological size due 

 to the development of post-occipital ridges. The 

 length is also increased at the anterior extremity 

 in very old specimens owing to occasional 

 ossification of the cartilage at the tips of the 

 nasals. 



In the case of the measurement of the length of the 

 nasals an element of uncertainty is often introduced by the 

 irregularity of the suture with the frontal. This error may be 

 either positive or negative. 



In the Zygomatic breadth, a negative (i.e. the measure- 

 ment obtained is too small) error is introduced by the spring 

 of the zygomatic arches. 



The Diastema is affected by the position of the roots ot 

 the anterior premolars which spread forward to a variable 

 degree. This error also is negative. 



The tooth-row measurement, which is taken on the 

 alveolus, is affected in the same way, though the error in this 

 case is positive (i.e. the result is too large); and also in old 

 skulls by actual absorption of the teeth when the sign is 

 negative. This is not very marked in most rats* but the genus 

 Rhinoscinrus (Sciuridae) may be cited as an extreme case. 



The specimens which have been measured, have been 

 selected as adult, those specimens which show no signs 

 whatever of wear on the molars having been rejected. 



The arithmetic mean error, the error of mean square or 

 Standard Deviation of Pearson have been calculated as also 

 the Coefficient of Variation. It will be noted that the 

 measurements in all cases, if plotted, form curves of a 

 symmetrical type, the arithmetic mean agreeing very closely 

 with the Median. 



In the case of the upper tooth-row I have not given the 

 standard deviation or the arithmetic mean error as the 

 measurement does not admit of sufficient accuracy to give 

 consistent results, the actual dimensions being very small and 

 the normal variation being apparently contained within very 

 narrow limits. 



Table I. Measurements of Epimys ratttts jarah 



(Bonhote) 



Adult males. 



Table II. Do. 



Adult females. 



