A leaf by any other name

The leaf-like segments of Schlumbergera, are parts of the
stem system.

What is a leaf?  For practical purposes, it might be any flat, photosynthetic plant organ.  Yet we know that there are certain “stems in leaf’s clothing” in the botanical world.  Cacti evolved in deserts, where leaves were a liability, and thick, succulent stems took over the job of photosynthesis.  Many cacti that have adapted as epiphytes in the tropics, such as Schlumbergera (Christmas cactus) and  Epiphyllum, however, have “reinvented leaves” by making their stem segments flat and thin. 

Many brown algae produce large, leaf-like fronds.
Line drawing from Allen & Gilbert, 1917, A textbook
 of botany. 

To be fully convincing as a leaf, a structure must not only be flat and photosynthetic, but also limited in size and shape (determinate), and produced in a regular pattern around a central stem.    Leaves also have a certain lifespan, after which they fall off of the plant, or sometimes remain as a dead skirt, as in Washingtonia palms.  New leaves are produced at the tips of stems that continue to elongate over time.  This would rule out leaf-like cacti, in which the flat segments are produced one from another like links in a chain.  They are parts of indeterminate, branching stem systems - stems in leaf’s clothing. 

Even within that more restrictive definition, flat photosynthetic appendages that are commonly referred to as leaves have evolved independently many times.  The leaf-like shape, not surprisingly, is nature’s most efficient light gathering antenna, and so has been reinvented over and over again. Many algae have adopted this highly successful growth form.  Kelp, for example, form underwater forests of long stems bearing many leaf-like fronds produced in sequence from an embryonic tip (an apical meristem).

The "leaves" of leafy liverworts, like this Lejeunea, are
flat extensions of the thallus.

The first flattened, photosynthetic structures to appear in land plants were the thalli of ancient liverworts.  A thallus is a plant body that is not clearly defined into organs like stems and leaves.  A thalloid liverwort is flat and photosynthetic, but grows and branches at its tip like a stem.  Some liverworts are called “leafy liverworts,” because their thalli are subdivided into small leaf-like segments with slender stem-like sections in-between.   Mosses are more convincingly leafy, with

determinate, leaf-like structures attached spirally around a stem,  but purists prefer to not call any bryophyte structures leaves because they evolved independently of the “true leaves” of other land plants.

 

Mosses, like this Barbula agraria, have distinct leaves produced one at a time by the stem apex.

Club mosses, like this Lycopdiella cernua
 from Florida, have small scale-like leaves
called microphylls.

However, the true leaves of vascular plants evolved at least twice from scratch, and were subsequently completely remodeled several times.   Early land plants had perennial creeping stems, called rhizomes, plus short-lived upright shoots adapted for gathering light and producing spores.   The early upright shoots were little more than green, forking stems, but competition for light soon forced them to evolve more efficient light-gathering structures.  In clubmosses (Lycophytes) the answer came in the form of flat but narrow leaves with a single vein of vascular tissue running through them.  They are referred to technically as microphylls, and are believed to have evolved as simple outgrowths of the surface tissues of ancient stems.  A more recent hypothesis is that microphylls evolved from sporangia that were “sterilized” and flattened.  Precursors of lycophytes produced numerous sporangia on short stalks along the sides of their upright leafless stems.  So converting some of them into leaves would have been a fairly simple adaptation.  In either case, leaves of lycophytes can grow in length, but cannot develop complex shapes or much breadth.

The fronds of ferns are
upright shoots flattened into
a leaf-like configuration.  From
Smith, 1955, Cryptogamic Botany.

The complex fronds of ferns, which bear sporangia on their.

The large complex leaves of ferns are called megaphylls.

lower surfaces, as well as conducting photosynthesis are  upright shoots that became leaf-like
through fine-branching and  flattening.  Such leaves are called megaphylls.  Megaphylls can be called leaves because they are produced sequentially at the tips of the ongoing rhizomes, have a definite size and shape, and fall off of the plant after one or a few seasons

The upright shoots of horsetails are equivalent to the fronds of ferns,
 but consist of repeated whorls of small leaf-like branchesand
elongate stem segments.
From Kerner & Oliver, 1904, The natural history of plants.

The upright shoots of horsetails, cousins of the ferns, evolved a little differently.  They too are determinate, photosynthetic, and spore-bearing, and are discarded after a defined period of time, but they remained stem-like with smaller leaf-like segments.  Though modern horsetails don’t have leaves, their earliest ancestors had short, fan-shaped leaves born in a circular arrangement at intervals along the upright shoots.   They evolved a unique, bamboo way of growth, in which stem segments elongate to extend the entire shoot quickly upward (see The first “bamboos,” 28 Mar, 2012). These smaller leaves are also called megaphylls, though each is equivalent to only part of the fern megaphyll.  Spores are produced, not on the leaves, but in specialized cones at the ends of upright shoots. 

  This is an
ancient horsetail ancestor called Lilpopia,
with small megaphylls, each equivalent
to just a small part of a fern frond.

Cycads have compound leaves
descended from the fronds of seed ferns.

The leaves of the cycad Bowenia are doubly compound, and
the most like ancient seed ferns.

The leaves of flowering plants, as well as cycads, are

Archaeopteris was an ancient spore-bearing
tree that may have been a precursor to
both seed ferns and conifers. These lateral
leafy systems may have evolved directly into
seed fern fronds or into branch systems of
small needle-like leaves in the conifers.

also considered to be full  megaphylls, having evolved from the large fern-like fronds of plants known as “seed ferns.”  Seed ferns evolved in much the same ways as ferns, though probably independently from leafless ancestors, and bear seeds and pollen sacs directly on the leaves.    In conifers and ginkgos, however, needle-like and fan-shaped  leaves may represent segments of ancient megaphylls that became more twig-like.  Short leafy shoots of these plants would then be evolutionarily equivalent to entire seed fern fronds.  The common ancestors of all seed plants, the progymnosperms, had leafy systems that were complex and frond-like, but not clearly defined as either determinate or indeterminate.   Thus both types of systems could have evolved from them. 

The leaves of conifers, such as this
Araucaria, are
simple, and flat or needle-like.

Angiosperm leaves, like this Tetrapanax,
can be large and complex.

Leaves in the eudicot family, Apiaceae, are typically
compound, and can be quite fern-like, as in this variety of
parsley.

The leaves of flowering plants, though evolving from seed-fern type ancestors, are extremely varied in structure.  Some are complexly branched, like their ancestors, others are small and simple, even scale-like in some cases.  Their extreme evolutionary plasticity demonstrates the innate potential for growth and complexity inherent to the original megaphylls.    Angiosperm leaves, moreover, develop in two different ways, in accordance with what we might call the “dicot model” and the “monocot model.” 

Dicotyledonous plants occur in several distinct clades, mostly in the Eudicot clade, but also in the more ancient Magnolid clade, and the most ancient clades of the ANITA grade (Amborella, Nymphaeales, and Austrobaileyaceae – another long story!).  In this developmental model, leaves begin as tiny peg-like primordia at the tips of the stems, after which they develop their characteristic shapes in miniature.  Complex, dissected, and irregular shapes develop through marginal meristems expanding locally at different rates.   After the shape has been formed (and in some climates after a period of dormancy within a protected terminal bud), the leaves expand two-dimensionally, increasing in size but retaining the shapes developed in their infancy.

In the eudicot, Liquidambar, leaves
develop their shape first in miniature,
then expand to their full size.

In the monocots, leaves being as hood-like primordial, with a basal sheath surrounding the apical meristem, then expand primarily through basal growth (see How the grass leaf got its stripes, 26 Jan 2012).  By growing only from the base, typical monocot leaves are long and strap-like and their veins of vascular tissue run parallel to one another.

The typical monocot leaf grows from the base,
 resulting in a strap-shaped structure and
parallel veins.
From Rost, et al., Plant Biology

All of these structures can be called leaves, though they develop in different ways.  Botanists will continue to use more precise technical terms for leaf-like structures that evolved independently.