The global cooling of the late Cenozoic was almost certainly accompanied by a change in circulation of the abyssal oceans. Today, the deep waters of the three major oceans blend in the Antarctic Circumpolar Current as they are continually transported from one ocean basin to the next one east. Pliocene time was marked by important paleoenvironmental changes, including the closing of the Panamanian Seaway at 4.5 Ma, pulses of ice rafting in the North Pacific at about 4.3 Ma, a strengthening of the Northern Hemisphere westerlies beginning 4.5 Ma, Mg/Ca evidence from Kerguelen of early Pliocene Southern Hemisphere cooling, and the million-year long cooling and drying trend that ended with the onset of major Northern Hemisphere glaciation at 2.6 Ma. All these phenomena indicate changes in the transport of heat through the Earth's climate systems. Studies of abyssal circulation based on inter-oceanic comparison of C-13 values in benthic foraminifera suggest a vigorous deep circulation throughout the early Pliocene; conversely, studies based on manganese nodule geochemistry suggest declining inter-oceanic exchange of deep waters then. Our studies of deep-sea drift deposits provide a third, physical, record of abyssal flow: anisotropy of magnetic susceptibility (AMS), or magnetic fabric, which is a measure of the alignment of the mineral grains in a sediment sample. The stronger the flow, the more pronounced the degree of magnetic anisotropy. Fabric data from the Feni and Gardar Drifts in the North Atlantic show a sudden change in strength from pronounced in the early Pliocene to very low in the late Pliocene. This change occurs at 2.6 Ma in Feni drift DSDP Site 610 (2417 m), and at 4.3 Ma in Gardar drift DSDP Site 611 (3220 m). Fabric measurements spanning the last 9.5 million years on an Antarctic Peninsula drift, recovered at ODP Site 1095 (3841 m), also show a decline in strength at about 4.5 Ma. AMS data from ODP Site 745 (4082 m) on the Kerguelen drift show a general decline in strength throughout the past 6.5 m.y. Taken together these data indicate a significant slowing of abyssal flow since at least the early Pliocene, and perhaps earlier. This implies a reduced ability for the deep ocean to transport heat, potentially a critical aspect of late Cenozoic high latitude cooling.