Northern Brazil contains remnants of Mesozoic flood basalts and hypabyssal rocks that were apparently emplaced during tectonism related to opening of the Atlantic Ocean. Analyses and new K-Ar ages reveal that this ∼700x250 km Maranhão province (5°–8°S) has low-Ti basalts (∼1.1 wt% TiO2) in the western part that range about 160 to 190 Ma, and high-Ti basalts (3.4–4.4 wt% TiO2) in the eastern part about 115–122 Ma. Low-Ti basalt compositions are less evolved and have a smaller range, Mg# 62-56, than the high-Ti basalts, Mg# 44–33. General characteristics of the least evolved members of low- and high-Ti groups include, respectively, Zr 100 and 250 ppm, Sr 225 and 475 ppm, Ba 200 and 500 ppm, Nb 10 and 26 ppm, Y 29 and 36 ppm, La/Yb(n) 4.2 and 8.8, where La(n) is 30 and 90. Overall compositions resemble the low- and high-Ti basaltic rocks of the Mesozoic Serra Geral (Paraná) province in southern Brazil. The Maranhão low-Ti basalts have more radiogenic Sr and Pb and higher δ18O than the high-Ti basalts. Respectively, low- vs high-Ti: ɛSr26−54 vs 15−18; 206Pb/204Pb=18.25–.78 vs 18.22–.24; and δ18O 8.9–12.6 vs 6.5–8.6. Nd isotopes overlap: ɛNd−1.6 to −3.8 vs −2.1 to −3. Ages, compositions, and isotopes indicate that the low- and high-Ti groups had independent parentages from enriched subcontinental mantle. However, both groups can be modeled from one source composition if low-Ti basalt isotopes reflect crustal contamination, and if the parentages for each group were picritic liquids that represent either higher (for low-Ti) or lower (for high-Ti) percentages of melting of that single source. When comparing Pb isotopes of Maranhão and Serra Geral high-Ti basalts (uncontaminated) to evaluate the DUPAL anomaly, Maranhão has Pb Δ7/4=4.6–11, and Pb Δ8/4=72–87; Serra Geral has Pb Δ7/4=10–13, and Pb Δ8/4=95–125. The small difference is not enough to conform to DUPAL contours, and is inconsistent with large-scale isotopic heterogeneity of mantle beneath Brazil prior to rifting of South America from Africa. Maranhão low-Ti magmas probably relate to the opening of central North Atlantic, and high-Ti magmas to the opening of equatorial Atlantic. The proposed greater percentage of source melting for low-Ti basalts may reflect a Triassic-Jurassic hotspot, while lesser melting for high-Ti magmas may relate to Cretaceous decompressional (rifting) melting.