Underdoped cuprates exhibit a normal-state pseudogap, and their spins and doped carriers tend to spatially separate into 1-or 2-D stripes. Some view these as central to superconductivity, others as peripheral and merely competing. Using La2−xSrxCu1−yZnyO4 we show that an oxygen isotope effect in Tc and in the superfluid density can be used to distinguish between the roles of stripes and pseudogap and also to detect the presence of impurity scattering. We conclude that stripes and pseudogap are distinct, and both compete and coexist with superconductivity.PACS numbers: 71.10. Hf, 74.25.Dw, 74.62.Dh, 74.72.Dn High-T c superconductors (HTS) remain a puzzle. Various correlated states have been identified in HTS including antiferromagnetism, the pseudogap[1], nanoscale spin-charge stripes [2] and, of course, superconductivity (SC). (Here we generalise"stripes" to include possible 2D checkerboard structures [3]). The pseudogap is a nodal energy gap of uncertain origin that appears in the normal-state (NS) density of states (DOS). Its effects can be observed in many physical properties [1,4]. Several opposing views are still current. One is that stripes play a central role [5], forming the pseudogap correlation [6] and/or mediating the SC pairing. Another is that the NS pseudogap arises from incoherent superconducting fluctuations which set in well above T c [7]. Another is that these states are independently competing [8]. Here stripes and pseudogap play a secondary role and SC is mediated by some other pairing boson. An unambiguous test of these opposing views is urgently needed. We show here that isotope effects provide such a test.The isotope exponent α(E) in a given property E is defined as α(E) = − (∆E/E)/(∆M/M ), where M is the isotopic mass and E may be T c , the SC gap parameter, ∆ 0 , the pseudogap energy scale, E g , or the superfluid density ρ s = λ −2 ab = µ 0 e 2 (n s /m * ab ). (λ ab is the in-plane London penetration depth, n s is the carrier density and m * ab is the effective electronic mass for in-plane transport). An isotope effect on T c was first discovered in 1950 by Allen et al. for Sn [9]. They found α(T c ) ≈ 0.5 ± 0.05 which provided the central clue for the role of phonons in pairing and led 7 years later to the BCS theory of SC [10].The situation with HTS is more complex. The oxygen isotope effect on T c was found[11] to be small, with α(T c ) ≈ 0.06. However, with decreasing doping the effect rises and eventually diverges as T c → 0[12, 13]. Surprisingly, an isotope effect was also found in the superfluid density [14] (and attempts were made to resolve this into a dominant isotope effect just in m * [15,16]). We will show that both of these unusual effects can be understood in terms of a normal-state pseudogap which competes with SC [17]. We also predict and confirm an isotope effect in ρ s induced by impurity scattering. The isotope effects in T c and ρ s are mapped as a function of doping in La 2−x Sr x Cu 1−y Zn y O 4 and we observe a canonical pseudogap behavior as well as a huge an...