Fourier transform infrared (FTIR) microspectroscopy has been used to estimate the spatial proteoglycan (PG) and collagen contents in articular cartilage (AC). However, it is not clear whether the results of FTIR analyses are consistent between different species. Our aim was to clarify how three different FTIR PG parameters in use, i.e., the integrated absorbance in the carbohydrate region, the carbohydrate/amide I ratio, and the second derivative peak at 1062 cm-1, can indicate the densitometrically assessed (reference method) spatial PG content in a sample set consisting of osteoarthritic human and bovine AC samples. The results show that all the parameters can accurately reflect the PG content, when the species are analyzed separately. When all samples are pooled, the correlation with the reference method is high (r=0.760, n=104) for the second derivative peak at 1062 cm-1 and is significantly lower (p<0.05) for the carbohydrate region (r=0.587, n=104) and for the carbohydrate/amide I ratio (r=0.579, n=104). Therefore, the analysis of the carbohydrate region may provide inconsistent results, if the cartilage samples from different species are in use. Based on the present study, second derivative analysis yields more consistent results for human and bovine cartilages.