A pot experiment with $P^{32}$ was carried out to investigate the soil phosphorus availability to four leguminous forage crops and three graminaceous, of black volcanic ash soil and red one. Soil phosphorus was extracted with 6 different extractants and also fractionated in Fe, Al and Ca phosphorus. The results were: 1) Soil phosphorus availability was in decreasing order of Italian rye grass${gg}$ soybean> cassia> corn> weeping love grass${gg}$ Korean lespedeza> Red clover and they might be grouped into three levels by A-value, over 1000, 200-500 and below 40 $p_{2}O_{5};kg/ha$. 2) The amount of various available phosphorus and phosoborus fraction in the black soil was higher than that in the red soil. No difference in phosphorus availabiliy to forage crops was shown between two soils. Therefore an extractant able to draw out similar amount of phosphorus from two soil will be suitable for determining the phosphorus availability index. 3) Two extractants, one extracting 20 ppm as maximum and the other extracting 100 ppm as minimum will be recommendable for determining the availability of phosphorus; the former for red clover and Lespedeza and the latter for others. Truog method may be good for the former but no appropriate method for the latter was found in the methods used. 4) T/R ratios of legumes were negatively correlated at 5% level with % phosphorus from fertilizer (% pdF). Legumes showed below 50 of % pdF over 5 of T/R ratio and over 80 of % pdF below 5 of T/R.