The value of transitional bilingual education (TBE) is discussed in the context of two competing theories of second language learning: (1) first language (L1) knowledge facilitates second language (L2) learning, and (2) the best way to learn English is to maximize time spent using it. First, the facilitation theory is examined in light of recent research and found to be flawed. Two significant national studies are shown to have findings contradicting it. The English time-on-task theory is then explored and found lacking, with evidence drawn from the same two studies. It is argued that the theory fails because it does not consider important mediating variables in program effectiveness. A third study is cited, one that suggests some variables explaining the success of TBE with less English-language time-on-task, including nature of time spent in an English language environment and psychological effects of providing some respite between exposures to the target language. Two new hypotheses are proposed, to be tested by research. The hypotheses are: (1) native language instruction should be minimal and used only in early instruction, and (2) teachers who are familiar with but not fluent in the child's native language are better teachers of limited-English-proficient students. In addition, a new experimental design is recommended. (MSE)