This urticlr qitunti'jes errors arising from vcirious integrcition methods and,from sampling density in thti numerical rstimcition i?f'cdorimrtric integrcils,e.g. tristimirlus ual-i4e.Y. Both dutu sumpling dtwsity and uarious numerical itrtegrtttion method errors will be distingitished. A test is presrntrd to yriantib sampling intrrual rryuirrmrnts f o r acrurutr cdorimetric calculations j b r various illuminunt.s. Cj I992 John Wile? & S o w , Inc.

Numerical melhodv ,f;w estimating cdnrimetric integrals using various data intervuls Lire reuie wed. Accurucy of Newton-Cotes and Gaussian numerical integrationulgo-rithn1.s ure compared. Tables of colorimetric dataJbr per-Jiwming Guuss-Legendre integrution are presented. Appropriati) numerical examples are presented and applicability t o general and spocial cases is discussed. Tho B-spline method of integration is presented. Errors of integrution are quantified and discussed. 0 1992 John