Eight albino mutants which behaved as single recessive were investigated on the linkage relation with several marker genes. Since the albinos are lethal at seedling stage, when the marker characters used were difficult or impossible to distinguish from the normal type at the stage, it is the only way to calculate the recombination value from the progeny test of them. Here we suppose that a and b are albino and marker genes respectively, two kinds of F_1 genotypes, AABb and AaBb, are produced from crosses between heterozygotes for albino gene (AaBB) and marker strains (AAbb) . In a heterozygous F_1 for the albino gene (AaBb), the frequencies of four gametes, AB, Ab, aB and ab, will be p/2, (1-p)/2, (1-p)/2 and p/2, respectively, where p is recombination value between a and b. Therefore, nine genotypes composing the four F_2 phenotypes and their expected and observed frequencies are as shown in Table 3. Now we think following five cases for estimating p. (1) F_2, 4 classes: When the marker gene used is possible to distinguish from the normal at the seedling stage, F_2 phenotypes are classified into four classes, AB, Ab, aB and ab. (2) F_2, 2 classes: When the marker gene used is impossible to distinguish from the normal at the seedling stage, F_2 phenotypes are classified into only two classes of survivals, AB and Ab. (3) F_3 lines from A . F_2: When the gene for endosperm character such as wx is used as marker, three genotypes of F_2, A BB, A_Bb and A_bb are possible to classify according to the segregation of wx in F_3 grains setting on the survived F_2 plants. (4) F_3 lines from B. F_2: F_2 individuals of AB phenotype are classified into two genotypes for albino, AAB_ and AaB_, by the progeny test at the seedling stage. (5) F_3 lines from bb F_2: F_2 individuals of Ab phenotype are classified into two genotypes for albino, AAbb and Aabb, by the progeny test at the seedling stage. On the five cases, estimation equations for determining p and formulas for calculating the amount of information per F_2 individual (i_p) are shown in Table 4. The standard error of p (SE_p) is given by the formula, SE_p =√<1/I_p>, where I_p is total amount of information (i_p×N). Using the above formulas, the recombination values (%) between the albino genes and marker genes (linkage group) were calculated as follows: al_<K-1>-wx(I) 7.1, al_<K-2>-nl_1(IX) 1.1, al_<K-3>-nl_I (IX) 34.6, al_<K-4>-lax (III) 13.3, al_<K-5>-lg (II) 33.7, al_<K-6>-nl_1 (IX) 35.5, al_<K-6>-ri (IX) 13.1, al_<K-7>-lg (II) 31.1, al_<K-8>-d_<18> (unknown) 10. 8. To compare the efficiency for estimation of p in the five cases, the relative efficiencies calculated from Ip (i_p × N) are shown in Fig. 1. In these cases, numbers of individuals (N) are calculated from the number of original F_2 (n) at n, 3 n/4, 3 n/4, (2+p^2) n/4 and (1-p^2) n/4 in the cases of (1), (2), (3), (4) and (5), respectively. A curve of the case (3) was omitted in the Figure because it was similar to that of the case (5). It is evident that the efficiency is extremely lower in the case (2) than in the other four cases, and that the efficiencies in the cases (4) and (5) were alike to that in the case (1) when p was about 0. 3 and became still higher than in the case (1) with the decrease of p. Therefore, it is well expected that there were some crosses found the linkage relation between the marker genes and the albino genes in the F_3 from survived F_2, nevertheless the significant linkage relations were not recognized in the survived F_2 (Table 7 and 9). As mentioned above, it is found that the lethal characters such as the albino are useful for linkage works in rice, notwithstanding there are some complications to maintain albino genes with heterozygous condition and to breed F_3 lines from the survived F_2 plants.