Although advanced gastrointestinal cancer is the most commonplace problem encountered by the medical oncologist, this group of diseases has proved exceedingly resistant to past chemotherapy efforts. 5-Fluorouracil (5-FU), accepted by some as standard treatment, had provided only infrequent, incomplete, and fleeting antitumor effects, which are probably more than counterbalanced by its gastrointestinal, mucocutaneous, and hematologic antihost effects. There is no evidence that any manipulation of route or schedule of administration provides any improvement in the therapeutic ratio of 5-FU. There is no evidence that this drug contributes to patient survival when used at any stage of any type of gastrointestinal carcinoma. The search for alternative single drugs to 5-FU has been disappointing. The nitrosoureas and Mitomycin C produce occasional regressions, but they do not match the meager effectiveness of 5-FU; and they, in addition, present the difficult problem of cumulative bone marrow suppression. Recent trials with combination regimens have given some indication that the long stalemate in chemotherapy of gastrointestinal cancer may be breaking. Substantial improvements in frequency of tumor regression have been recorded for gastric carcinoma with combinations of 5-FU and BCNU, 5-FU and methyl CCNU, and 5-FU, Mitomycin C, and cytosine arabinoside; for colorectal carcinoma, with the combination of 5-FU, methyl CCNU, and vincristine; and for carcinoid tumors and islet cell carcinomas, with the combination of 5-FU and Streptozotocin. There are also suggestion that such combination chemotherapy with response rates in the 30 to 50% range may produce increased survival when compared to the untreated patient and patients treated with single-drug regimens. While the accomplishments of chemotherapy for the gastrointestinal cancer patient remain less than spectacular there is nevertheless realistic hope that a respectable contribution can now be made to multidisciplinary efforts applied at a stage of disease with minimal tumor burden.