Current research on adult second language learning suggests that study is necessary to separate interference variables from intralinguistic variables and to determine the relative difficulty of structures in each group. This research on the acquisition of the English relative clause was specifically designed to test the transfer hypothesis specified by Lado. In the first portion of the study the role of interference is investigated; in the second part, a definition of a natural difficulty ordering for the stages of relative clause acquisition is proposed. Grammaticality judgments were elicited by presenting speakers of several different languages with a written list of 36 sentences, each containing a relative clause structure. Subjects were asked to judge them as structurally correct or incorrect. The main question was whether it is language group, sentence type, or a combination of the two that is more predictive of a learner's performance. Results show that sentence type rather than native language background is the most reliable predictor of error. An attempt is made in the second part of the study to ascertain if there exists a natural difficulty ordering among the various types of relative clause structures. If such an ordering is found to exist, it might be determined to what extent it reflects the acquisition strategies employed by first language learners of English. (AMH)