The nonminimal pure spinor formalism for the superstring is used to prove two new multiloop theorems which are related to recent higher-derivative R 4 conjectures of Green, Russo, and Vanhove. The first theorem states that when 0 < n < 12, @ n R 4 terms in the Type II effective action do not receive perturbative contributions above n=2 loops. The second theorem states that when n 8, perturbative contributions to @ n R 4 terms in the IIA and IIB effective actions coincide. As shown by Green, Russo, and Vanhove, these results suggest that d 4 N 8 supergravity is ultraviolet finite up to eight loops. DOI: 10.1103/PhysRevLett.98.211601 PACS numbers: 11.25.Db Introduction.-The main success of superstring theory is that it provides a consistent quantum theory of gravity which replaces general relativity at small distances. It is therefore important to know how graviton scattering at high energies differs in superstring theory from general relativity. For scattering amplitudes involving two incoming and two outgoing gravitons, the lowest-order deviation from general relativity comes from R 4 terms in the superstring effective action [1], and higher-order deviations come from @ n R 4 terms where @ n denotes n spacetime derivatives, R denotes the Riemann tensor, and Lorentz indices are suppressed. The structure of these @ n R 4 terms is believed to be tightly constrained by duality symmetries of the superstring [2]; however, these duality symmetries are difficult to prove since they involve nonperturbative effects. It is therefore important to compute @ n R 4 terms in perturbative superstring theory, both for studying deviations from general relativity and for testing the nonperturbative duality symmetries.A further motivation for studying the structure of @ n R 4 terms in Type II superstring theory is that they provide information about the maximally supersymmetric version of gravity in four dimensions which is called d 4 N 8 supergravity. In the early days of supersymmetry, d 4 N 8 supergravity was conjectured to be a consistent finite theory of gravity. It was later realized that d 4 N 8 supergravity probably has ultraviolet divergences which are eliminated only after including the massive states of superstring theory. However, the existence of these ultraviolet divergences in d 4 N 8 supergravity has never been proven, and it was recently shown by explicit computation that they are absent up to three loops [3]. Furthermore, it was argued in [2,4,5] that finiteness properties of d 4 N 8 supergravity are related to nonrenormalization theorems of @ n R 4 terms. For example, using the results described here that @ n R 4 terms do not get contributions above n=2 loops when n < 12, it was deduced in [5] that d 4 N 8 supergravity is ultraviolet finite up to 8 loops.Although there are several prescriptions available for computing superstring amplitudes, the most efficient pre-