How arsenic (As) is transported in phloem remains unknown. To help answer this question, we quantified the chemical species of As in phloem and xylem exudates of castor bean (Ricinus communis) exposed to arsenate [As(V)], arsenite [As(III)], monomethylarsonic acid [MMA(V)], or dimethylarsinic acid. In the As(V)- and As(III)-exposed plants, As(V) was the main species in xylem exudate (55%-83%) whereas As(III) predominated in phloem exudate (70% 94%). The ratio of As concentrations in phloem to xylem exudate varied from 0.7 to 3.9. Analyses of phloem exudates using high-resolution inductively coupled plasma-mass spectrometry and accurate mass electrospray mass spectrometry coupled to high-performance liquid chromatography identified high concentrations of reduced and oxidized glutathione and some oxidized phytochelatin, but no As(III)-thiol complexes. It is thought that As(III)-thiol complexes would not be stable in the alkaline conditions of phloem sap. Small concentrations of oxidized glutathione and oxidized phytochelatin were found in xylem exudate, where there was also no evidence of As(III)-thiol complexes. MMA(V) was partially reduced to MMA(III) in roots, but only MMA(V) was found in xylem and phloem exudate. Despite the smallest uptake among the four As species supplied to plants, dimethylarsinic acid was most efficiently transported in both xylem and phloem, and its phloem concentration was 3.2 times that in xylem. Our results show that free inorganic As, mainly As(III), was transported in the phloem of castor bean exposed to either As(V) or As(III), and that methylated As species were more mobile than inorganic As in the phloem.