Although normal isotropic hummocky cross-stratification is commonly interpreted to be the deposit of large-scale ripples, there are many reasons why this may not usually be the case. These reasons include: (i) that the stratification produced by large-scale ripples does not particularly look like isotropic hummocky cross-stratification; (ii) that it is difficult reconciling the abundance of HCS with the restricted hydraulic stability of large-scale ripples in silt to fine sand (i.e. the grain sizes in which hummocky cross-stratification is usually found); (iii) that the distribution of hummocky cross-stratification within ancient storm beds is not the distribution that would be expected from large-scale ripples; (iv) that the flows calculated to have formed ancient examples of hummocky cross-stratification would be expected to generate an upper stage plane bed rather than ripples; and (v) that it is difficult to explain why large-scale ripples would predominate in the proximal parts of storm beds when modern storm flows commonly exceed the threshold for entrainment. In contrast to the various hypotheses which propose that isotropic hummocky cross-stratification is generated by ripples, an alternative hypothesis which suggests that it is generated by instabilities, does seem to adequately explain the origin of hummocky cross-stratification. However, it is difficult to accept this hypothesis given that the origin of the proposed instabilities is unproven. These conclusions highlight the continued uncertainty regarding the process, which generates hummocky cross-stratification.