The first cytoplasmic male sterile (CMS) lines in cotton to have practical utility were developed by Vesta Meyer at Stoneville, Mississippi from the introgression of the cytoplasm of Gossypium harknessii into G. hirsutum. Nearly all subsequent work with CMS/restoration in cotton has utilized lines derived from Meyer's work. However, evidence has accumulated in recent years that the G. harknessii cytoplasm causes a 7-10% reduction in yield, hence, apparent heterosis in hybrids with the D(2) cytoplasm is less than expected. In 1992 Stewart reported the development of another cytoplasmic male-sterile (CMS) restorer system in upland cotton based upon the cytoplasm of G. trilobum (D(8)). The objective of this work was to describe the morphological characteristics of the new D8-CMS line. Light and electron microscopy were used to examine anther ultrastructure. Anthers of D8 heterozygotes and ARK8518 showed normal microsporogensis, whereas the tapetum and sporogenous tissues of D8-CMS anthers abort and disintegrate before meiosis. Pollen viability counts, determined by the FCR method, showed that sterility expression was not at the microspore level. Expression of D8-CMS decreased flower size including the corolla, pistil, calyx, style and ovary, compared to the ARK8518 B line. The male sterile flowers also had 18% fewer anthers per flower than ARK8518. However, bract size and number of ovules per boll were equivalent to ARK8518, and fully expanded bolls dimensions were larger. D8-CMS produced longer, finer fiber than ARK8518 and strength was the same. Arealometer tests indicated that perimeter and wall thickness was less in D8-CMS. At Fayetteville, AR boll weight of the D8-CMS was less than ARK8518, however, in yield trials at Keiser, AR, seedcotton yields per acre of the two were equivalent.