Traditionally, a microphyll is "an appendage supplied by a single, unbranched vein". Despite their name, microphylls are not always microscopic; those of Isoetes (quillworts) reach centimetres in length, and the extinct Lepidodendron bore microphylls over a metre long. In the classical concept of a microphyll, this vein emerges from the protostele, without leaving a gap.
Megaphylls, in contrast, leave a "leaf gap" when they depart the stele, with some vascular strands leaving to supply the leaf, and the other strands closing up-stem of the divergence. Megaphylls are characterised by multiple venation. By this definition, the whisk ferns (psilopsids), club mosses (lycopods) and horsetails (sphenopsids) have microphylls, as all extant individuals only bear a single vascular trace in each leaf.
The "Enation theory" of microphyll development posits that small outgrowth, or enations, developed from the side of early axes (such as those found in the Zosterophylls) . Outgrowths of the protostele later emerged towards the enations (as in Asteroxylon), and eventually continued to grow fully into the leaf to form the mid-vein (such as in Baragwanathia). The fossil record appears to display these traits in this order, but this may be a coincidence, as the record is very incomplete: microphylls may have originated by the reduction of a webbed telome branch.
By contrast, megaphylls appear to have originated from dichotomous branches. These first overlapped (or "overtopped") one another, and eventually developed "webbing" and evolved into gradually more leaf-like structures. So megaphylls, by this "teleome theory", are composed of a group of webbed branches. Hence the "leaf gap" left where the leaf's vascular bundle leaves that of the main branch resembles two axes splitting.
Unfortunately, the simplistic models described above do not hold true for all organisms. Some genera of ferns display complex leaves which are attached to the pseudostele by an outgrowth of the vascular bundle, leaving no leaf gap. Horsetails (Equisetum) bear only a single vein, and appear for all the world to be microphyllous; however, in the light of the fossil record, we conclude that their forbears bore leaves with complex venation, and the current state is a result of secondary simplification. Even some conifer needles bear only a single vascular trace, but again, this evolved as a secondary simplification from a multi-veined leaf.
An interesting case is that of Psilotum, which bears a protostele, and enations devoid of vascular tissue. Consequently, Psilotum was long thought to be a "living fossil" of close affinity to early land plants (Rhyniophytes). However, genetic analysis has shown Psilotum to be a reduced fern.
We have problems in the other direction, too. The lycopods are the only clade accepted to have evolved microphylls de novo, instead of by reduction from a megaphyll, but even they throw up exceptions: some Selaginella species have a complex venation instead of microphylls.
To make matters worse, there is also some debate about whether leaf gaps are unique and/or common to megaphyllous organisms.
This ambiguity leaves it difficult to distinguish between two competing hypotheses: that microphylls evolved via the reduction of megaphylls, and that they evolved independently, from enations. Taxonomically, the terms are perhaps better left undistinguished until more is known of their origins - perhaps the term "leaf" is more appropriate.