Abstract Resolution of a complex Hermitian matrix derived from monthly mean 850-hPa wind fields during the summer season (June–August) from 1968 to 2004 revealed four different modes of East Asian summer monsoon (EASM) variability. The leading EASM mode, accounting for 19.6% of the variance, is characterized by two different modes (M11 and M12) or their combination. Both portray a closed cyclonic or anticyclonic circulation anomaly over the western North Pacific (WNP), South China Sea (SCS), and southeastern China; corresponding anomalous geopotential height fields show a wave train structure from the WNP across Japan, the Okhotsk Sea, and Alaska to North America. Thus, the leading EASM mode characterizes the teleconnection pattern of the WNP-EASM. The correlation between M11 (M12) and the dynamic index for the WNP-EASM is 0.85 (0.51). M11 has leading spectral peaks at 15 and 3 yr, whereas M12 displays a predominant peak at 2 yr. It is found that M11 has interdecadal variations, with the transition years being circa 1973 and 1989, respectively. M11 is closely related to air–sea interactions in the SCS and the northwestern Pacific, and its association with the convective heat source over the northwestern Pacific is secondary. In contrast, M12 is closely related to the tropical convective heat source rather than tropical western Pacific sea surface temperature (SST). The second EASM mode, accounting for 12.8% of the variance, is identified and characterized by two distinct and alternating modes or their linear combination (M21 and M22). One mode (M21) closely relates to the dual blocking high pattern detected in anomalous sea level pressure (SLP) and 500-hPa geopotential heights over the Ural Mountains and the Okhotsk Sea. The other (M22) corresponds to a dipole blocking anomaly in anomalous SLP and geopotential heights, with opposing anomalous centers in the south of Japan and the Korean peninsula, and the area between Lake Baikal and the Okhotsk Sea. M22 shows significant correlations with summer mean rainfall in southern and southeastern China. Thus, a single index of EASM is inappropriate for investigating and predicting the EASM.