Müller (1904) stated that the axillary artery in the case of Adachi's C-type brachial plexus (AxC) might be derived from the 9th segmental artery. Yamada (1967) named a type of the subscapular artery (Sbs) "the superficial subscapular artery" which arose from the normal axillary artery (Ax), crossed over the medial cord of the brachial plexus and then gave off the lateral thoracic artery (TL). He considered that it might be derived from TL and develop to form AxC by compensating the less developed normal Ax. We reexamined the courses of Sbs and Ax and distinguished three types (S-, I-, and P-type) of Sbs according to their origin and course. Then we stated that the mechanism of formation of Sbs variations could be explained by the combination between the three stem parts and the common peripheral arterial network (Sbs system) (Aizawa et al. 1995). Therefore, the purpose of this study was to justify the validity of Müller (1904) and Yamada (1967) and to clarify the origin of AxC by applying the concept of Sbs system. The materials were 15 cases of AxC and 7 cases of incomplete AxC (AxC). The results were as follows. 1) The course of AxC was divided into four parts. 2) Two types of AxC were discerned according to the course against the nerve bundle communicating from the medial cord to the radial nerve (FM-R). They are the type-1 AxC which does not pass between the FM-R and the radial nerve, and the type-2 AxC which dose pass between them. 3) The first part included the branching points of the thoracoacromial artery in all cases and the superior superficial brachial artery (BSS) in 8 cases. The BSS passed between C7 and C8 of the roots of Ansa pectoralis (50%) in about the same manner as BSS from the normal axillary artery (Ax). On the other hand, the point where Ax or AxC penetrated the ventral stratum of the brachial plexus was examined in 156 cases. The data except those of the AxC cases displayed a symmetrical distribution having a sharp peak in C7-C8 (79.5%) and were not compatible with the incidence of AxC penetrating lower than Th1 (7.7%). Therefore, it was difficult to conclude that the first part of AxC was derived from the 9th segmental artery. 4) The second part crossed over the medial cord and gave off TL in almost all the cases. Therefore, this part was considered to include the S-point where the S-type Sbs system (Yamada's superficial subscapular artery) arose and to be derived from TL. 5) From the S-point, while the S-type Sbs system immediately ran down to the deep region of the axilla, AxC traversed the axilla passing in front of the thoracodorsal nerve to reach the point where AxC was sandwiched between the ventral and the dorsal stratum of the brachial plexus. Therefore, the following course from the S-point of AxC (the third part) was different from that of the S-type Sbs system. From the third part of AxC, the I-type Sbs system arose in 15 cases, and both the subscapular branch (RS: *) and the branch to the coracobrachial muscle (CB) were often given off. They were the same branches as those which arose from the I-point of normal Ax, and type-2 AxC passed between FM-R and the radial nerve in this part. Therefore, it was considered that the third part included the I-point of the normal Ax and, moreover, AxC recovered the normal course of Ax at the I-point. 6) The fourth part of AxC included the P-point where the P-type Sbs system branched off from AxC in 7 cases. The course of the fourth part of AxC had exactly the same course as that of normal Ax. 7) It was elucidated that the first part, the distal half of the third part, and the fourth part of AxC were exactly the same as normal Ax, the second part was derived from TL, and the proximal half of the third part from the S-point to the I-point was unique in AxC. Recently, however, the reverse course of the unique part of AxC has appeared as the deep lateral thoracic artery (TLp) (Aizawa et al. 1995) in rare cases. 8) In co