For the first ~40 million years (Myr) since their appearance around (or before) 425 Myr ago, throughout the Early Devonian, the early euphyllophytes were part of what has been called the "stick-world" (Tomescu and Whitewoods 2024) -- a leafless world that predates the evolution of leaves, in which terrestrial landscapes were populated by vascular plants whose sporophytes were simple branching systems consisting of segments with axial symmetry (axes). Leaves evolved later, starting in the Middle Devonian, in such morphologically simple euphyllophytes that by the end of the Devonian ~360 Myr ago, several leaf-bearing lineages were present. Leaf evolution has been a mainstay among the major areas of debate and investigation in plant morphology since its origin as a discipline, and our understanding of this evolutionary process has been tremendously influenced by three articles published in IJPS in the late 2000s (Sanders et al. 2007, 2009; Galtier 2010).
When the Sanders et al. (2007, 2009) and Galtier (2010) articles were published, the independent origin of leaves in the two clades of vascular plants (lycophytes and euphyllophytes) had been settled only recently (Kenrick and Crane 1997; Friedman et al. 2004) and was due in no small part to advances in understanding the deep tracheophyte fossil record. However, debates persisted about whether leaves evolved independently among the euphyllophytes and if so, how many times the leaves had evolved independently. This was in part because of an incomplete understanding of the deep phylogenetic relationships in this >425-Myr-old clade (especially the relationships among the denizens of Early Devonian stick-world and those between the Early Devonian plants and the younger Devonian euphyllophytes), despite apparent resolution achieved with studies that exclusively sampled living euphyllophytes.
Another major reason for this ongoing debate about euphyllophyte leaf evolution was that leaves evolved in this clade very long ago (>350 Myr). This tremendous time gap could not be easily bridged by comparative studies of leaf developmental regulation in living representatives of the clade. This then left the fossil record as the only other source of data that could provide answers to questions about leaf evolution.
However, the debates on leaf evolution saw a dearth of attempts at deliberate and thorough integration of structural data from the fossil record until the late 2000s, when three IJPS articles published in close succession changed the status quo by outlining an empirical approach (Sanders et al. 2007) and explicitly integrating data from fossil plants in an evo-devo framework to address euphyllophyte leaf evolution (Sanders et al. 2009; Galtier 2010). In doing this, they paved the way toward integration of the fossil record into a synthetic understanding of leaf evolution in the clade.
The starting point that Sanders et al. and Galtier were to build their studies on had been defined one decade prior to their articles, when Kenrick and Crane (1997) published a seminal study that established an empirically supported baseline for discussions of deep phylogeny and morphological evolution across the entire tracheophyte clade. In it, Kenrick and Crane proposed that the leaves of different euphyllophyte lineages evolved from an ancestral precursor structure of the leafless euphyllophyte denizens of stick-world: a lateral branch. However, as their study focused chiefly on phylogeny, they did not discuss any specifics about the features that may have distinguished the branching system precursor of leaves (i.e., the level of organization at which such leaf precursors might have been homologous) or the mode of evolution of leaves from such precursors, two topics that shape the very core of our understanding of leaf evolution.
Before Kenrick and Crane, the most direct and detailed pre-twenty-first century attempt at bringing the fossil record to bear on leaf evolution in euphyllophytes was by Jean Galtier, one of the deans of coal-ball paleobotany. Foreshadowing his 2010 IJPS article, Galtier undertook a detailed anatomical study of 340-350-Myr-old early seed plants (pteridosperms) and pteridophyte-grade plants of the Early Carboniferous (Mississippian) and compared them to older relatives that populated the Devonian stick-world (Galtier 1981). His data on the external morphology and vascularization of lateral appendages led Galtier to suggest that leaves evolved asynchronously in seed plants and filicalean ferns (with leaf evolution by the end of the Devonian in the former preceding leaf evolution in the latter) and that they evolved by different processes in the two lineages. This hypothesis implied the existence of different pathways taken by different plant lineages to evolve leaves from the simple branching trusses of stick-world denizens. Equally important, Galtier's study foreshadowed, albeit without explicitly prescribing them, the elements of an empirical approach to integrating fossils in understanding leaf evolution that fit squarely, albeit not declaredly, in the evo-devo paradigm. By breaking down leaf identity into discrete characters (laminar morphology, bilateral symmetry, dorsiventral [i.e., adaxial-abaxial] polarity) and documenting the presence or absence of each of those characters in fossils that could be placed taxonomically and dated, this approach could illuminate both tempo and mode in leaf evolution.
The integration of structural data from fossils to inform the evolution of plant development had also been a focus of Gar Rothwell (e.g., Rothwell 1987, 1992), another paleobotanist who had honed his skills at the school of coal-ball paleobotany. Rothwell's evo-devo interests, which had been recently boosted by his collaboration with Simcha Lev-Yadun on the evolution of cambial regulation by auxin (Rothwell and Lev-Yadun 2005), gained additional impetus with the arrival in his lab at Ohio University of Heather Sanders, who became interested in leaf evolution. Their collaboration with plant molecular developmental biologist Sarah Wyatt resulted in the Sanders et al. (2007) article published in IJPS on the connections between fossils, development, and evolution. In it, they introduced the concept of fingerprints of developmental regulation, and with it, they explicitly laid out the logic and formulated the empirical basis for integrating data from fossils to inform the evolution of plant development. Their approach turned the causal relationship between factors that regulate development and their phenotypic expression on its head, proposing to track the structural fingerprints through the fossil record and to use their patterns of occurrence and association over geologic time and across phylogenetic space to make inferences on the evolution of the developmental regulatory programs. This blueprint for using fossils as direct contributors to shaping hypotheses in the modern evo-devo paradigm sparked new synergy between paleobotany and evolutionary biology.
One immediate ramification of these advances was another article published in IJPS by the same team. In it, Sanders et al. (2009) looked at the evolution of leaves, following in Galtier's (1981) footsteps. Like Galtier but this time explicitly, they split leaf identity into discrete phenotypic characters that were used as fingerprints to reconstruct sequences of structural alterations reflected by fossil phenotypes in the fern and seed plant lineages. Sanders et al. used these sequences in turn to reconstruct sequences of stepwise developmental changes in the evolution of stem-leaf organography in the two lineages. While Galtier's inferences about leaf evolution had been tentative, in part because of the fragmentary state of fossils and some taxonomic uncertainty brought about by it, Sanders et al. were now bringing into the equation data from Elkinsia, the oldest seed plant and the representative of the pteridosperm grade, reconstructed as a whole plant (Serbet and Rothwell 1992). They were able to document dissimilar sequences in the origination of characters (growth determinacy, bilateral symmetry, pinnate architecture) and therefore independent evolution of developmental mechanisms for lateral organs in seed plants as compared to ferns. Their findings validated Galtier's hypothesis and contributed additional evidence supporting multiple independent origins of leaves in the euphyllophyte clade.
Less than one year later, Galtier (2010) expanded considerably on his 1981 IJPS article and Sanders et al.'s (2009) results, producing the most comprehensive review of leaf precursor structures and early leaves in euphyllophytes at the time. Aimed at addressing the origin and evolution of leaves in the clade, Galtier's review covered the Devonian and Early Carboniferous (Mississippian), thus spanning a consequential world change -- the transition from stick-world to the world of leafy plants. The review examined the major recognized plant groups of that interval (e.g., cladoxylopsids, progymnosperms) and the less well-known or poorly represented ones, such as stenokolealeans and rhacophytaleans. In contrast to his 1981 work, here Galtier explicitly provided an exhaustive list of the "diagnostic features" of leaves, whose presence he evaluated throughout the Devonian (Mississippian) fossil record. These structural characters included both morphological and anatomical features, and Galtier put particular emphasis on the latter as more accurate indicators of development. His detailed analysis provided evidence for far-reaching inferences on euphyllophyte leaf evolution, such as (1) leaf precursor structures and bona fide leaves evolved independently several times in the clade, between the Middle Devonian and early Carboniferous; (2) specific evidence for independent origins of leaves in archaeopteridalean progymnosperms, in seed plants, and two or more independent origins among the ferns (i.e., all the nonprogymnosperm and nonspermatophyte euphyllophytes); and moreover, (3) distinct patterns of acquisition of specific characters in different plant groups, thus providing novel, detailed insights into the mode of leaf evolution.
At the time, the studies by Sanders et al. and Galtier were not majorly acknowledged by the evo-devo community, which focuses mainly on documenting and comparing developmental regulatory mechanisms in living species, and they have been often overlooked since, but the articles were game changers nevertheless. Chief among their inferences, they demonstrated with unequivocal evidence that different euphyllophyte lineages followed different paths to escape the physiological constraints of stick-world by evolving leaves.
The far-reaching impact of the Sanders et al. and Galtier studies was recently emphasized by one of their most direct ramifications. Building on their approach and data and incorporating a wealth of fossil discoveries from the intervening decade and a half, Tomescu and Whitewoods (2024) found additional support for the conclusions of the three studies and uncovered further patterns in early leaf evolution. Tomescu and Whitewoods showed that leaf identity traits evolved independently across lineages, implying that the corresponding regulatory pathways were also combined independently into the integrated networks that regulate leaf development in different extant lineages. Furthermore, the approach and data of Sanders et al. and Galtier allowed Tomescu and Whitewoods to revisit the starting point of the three studies and address the conundrum of the hypothesized leaf precursor structures proposed by Kenrick and Crane (1997). For these, Tomescu and Whitewood (2024) proposed that to escape stick-world, different euphyllophyte lineages evolved leaves independently from a shared type of lateral branches characterized by regular taxis (arrangement).
Beyond their specific inferences, the studies by Sanders et al. and Galtier had tremendous epistemic implications for evolutionary biology; building on each other, they cemented the role of fossils as major, irreplaceable contributors to studies reconstructing the evolution of development and established a way to integrate them in such studies. Proof of their long-lasting impact is evident in the many ramifications these studies have generated over time (e.g., Rothwell et al. 2014; Hetherington et al. 2016; Tomescu et al. 2017; Rothwell and Tomescu 2021; Tomescu and Rothwell 2022; Pfeiler and Tomescu 2023; Hetherington 2024).