The conformation of native double helical DNA is well-known, but it is possible that small regions occur within native DNA, undetectable by X-ray diffraction methods, which have different conformations. Model structures are the synthetic deoxypolynucleotides of defined sequence. Under the conditions used, DNA, poly d(A-T) • poly d(A-T), and poly d(T-G) • poly d(C-A) can all give similar X-ray diffraction patterns, whereas poly dA • poly dT, poly dI • poly dC, poly dG • poly dC, and poly d(T-C) • poly d(G-A) clearly differ from DNA. This led to the tentative hypothesis that those DNA's in which all purines are in one strand and all pyrimidines in the other differ in structure from those (such as native DNA) in which purines and pyrimidines alternate or are irregular. We now find that poly d(I-C) • poly d(I-C) does not fit the hypothesis and is a most unusual structure, having seven or eight base pairs per turn. Both molecular model building and circular dichroism studies suggest that it is a left-handed helix. A number of purified tRNA's have been crystallized. We have obtained, from unfractionated tRNA, crystalline “powder” X-ray diffraction patterns showing rings and spots to about 20 A resolution. It is not clear whether cocrystallization has occurred, or whether there is fractional crystallization, though preliminary evidence favors the latter. Determination of the structure of crystalline tRNA has many features in common with protein crystallography, but there are a number of distinct differences.