In this talk, I will briefly review single-axon data, mainly for NHP, in the context of multi-scale distributed processing. In contrast with rapid advances in structural microcircuitry, progress at the network level has been slow.  Anatomical investigation of long-distance connections, especially in NHP, has been heavily skewed to retrograde tracing, and the small set of published reconstructions for cortical extrinsic projections have been limited to several mm at the distal arborization. This has supported global characterizations according to laminar and spatial features (divergent, layer 1 targeting or patchy, layer 4 targeting), but there are further categories and other connections that need to be considered.  Ultimately, whole neuron visualization, as is becoming more feasible in rodents, is necessary, where we can evaluate the complete intrinsic collateralization together with the complete extrinsic arborization, in their dynamic coordination, and in relation to identified neurons.  A key premise is that we are facing a diversity that is pronounced and functionally important. A recent study of mouse visual cortical projections concludes “…the dominant mode of intracortical information transfer is not… ‘one neuron-one target’…[but sharing] across subsets of target areas, and thus concurrently [contributing] to multiple functional pathways” (Y. Han et al., 2018).  Other results are consistent with this view; for example, single-axon reconstructions of thalamocortical axons in mouse (e.g., F. Clasca et al., 2012; E. Kuramoto et al., 2017) and of cortical connections from mouse barrel cortex (T. Yamashita et al, 2018); for NHP, Meynert cell collateralization (DMV Weisenhorn et al., 1995) and collateralized feedback connections (Rockland and Knutson, 2000; Suzuki et al., 2000); and in cat, heterogeneous intrinsic collateralization (KA Martin et al., 2014). These data invite revised perspectives on laminar, columnar, hierarchical, and other aspects of cortical organization.