Pre-Carpels from the Middle Triassic of Spain

In stark contrast to the multitude of hypotheses on carpel evolution, there is little fossil evidence testing these hypotheses. The recent discovery of angiosperms from the Early Jurassic makes the search for precursors of angiosperm carpels in the Triassic more promising. Our light microscopic and SEM observations on Combina gen. nov., a cone-like organ from the Middle Triassic of Spain, indicate that its lateral unit includes an axillary anatropous ovule and a subtending bract, and the latter almost fully encloses the former. Such an observation not only favors one of the theoretical predictions but also makes some Mesozoic gymnosperms (especially conifers and Combina) comparable to some angiosperms. Combina gen. nov. appears to be an important chimeric fossil plant that may complete the evidence chain of the origin of carpels in geological history, partially narrowing the gap between angiosperms and gymnosperms.


Introduction
The origin of angiosperms and their relationship with other seed plants have been the foci of botanical debates for a long time [1,2]. Carpels (the basic units of gynoecia in angiosperms) are idiosyncratic to angiosperms [3]. According to the traditional theory, a carpel results from the longitudinal folding and enrolling of a megasporophyll that bears ovules along its margins [4][5][6][7]. This hypothesis sounded rational, especially when Goethe's dictum "Alles ist Blatt" was taken into consideration [4]. However, it has been discarded since the APG system came into existence [8], and the APG system cannot give a plausible morphological interpretation for carpel homology [9]. Thus, plant systematics has entered a dead end: no widely accepted interpretation for the origin and homology of carpels is given, leaving many botanical questions unanswered. Therefore, using a fossil reproductive organ morphologically intermediate between angiosperms and gymnosperms to sift one hypothesis out of many is of crucial importance to move plant systematics beyond this debate. Although challenged by three groups of authors [10][11][12], Nanjinganthus, based on over 200 specimens of flowers, remains robust as an angiosperm from the Early Jurassic, since these challengers could not reach a consensus on the definition of angiosperms among themselves [13]. The Early Jurassic age of Nanjinganthus [13][14][15] suggests that the Triassic is a promising period for the search for a carpel precursor. Here, we report a new cone-like reproductive organ, Combina gen. nov. (Figures 1 and 2), from the Anisian (the lower Middle Triassic, >242 Ma) of Spain. In contrast to the seed-scale-bract-complex (SSBC) frequently seen in conifer cones, each lateral unit in Combina gen. nov. comprises an anatropous ovule in the axil of a bract that folds longitudinally and almost fully encloses the ovule. Such a configuration demonstrates a certain resemblance to both SSBCs in some Mesozoic "conifers" and carpels in some extant basal angiosperms (e.g., Illicium, Michelia) [16,17], suggesting a possible common ancestor shared by some angiosperms and Mesozoic "conifers". Its unique morphology and Triassic age make Combina gen. nov. one

Diagnosis:
Reproductive organ, cone-like and cylindrical, with helically arranged lateral units. Each lateral unit including an axillary ovule and a subtending bract. Ovule anatropous, attached to the organ axis. Bract longitudinally folded, with a ventral longitudinal suture.
Etymology: Combina from the Latin word "combinare", meaning "unite two things together", since the fossil combines the characteristics of angiosperms and conifers. ( Figure 1) Diagnosis: In addition to the genus diagnosis: organ with at least 13 lateral units helically arranged. Lateral units decreasing distally in size. Bracts elliptical to ovate in shape, up to 10 mm long. Ovules attached to the organ axis, up to 8 mm long, with a smooth outline and an asymmetrical base.
Description: The organ is cylindrical in form, 38 mm long and 16 mm wide, including more than 13 lateral units (Figure 1a). The lateral units are helically arranged along the organ axis, decreasing distally in size ( Figure 1a). The lateral units diverge from the organ axis at angles between 27° and 66° ( Figure 1a). Each lateral unit comprises a subtending bract and an axillary ovule (Figure 1b,c, g-j). The bract is 7 to 10 mm long, elliptical to ovate in shape, almost fully enclosing the ovule from bottom and laterals, leaving a ventral gaping suture (Figure 1a

Discussion
The general morphology of Combina appears to be that of a coniferous cone, in which  ( Figure 1) Diagnosis: In addition to the genus diagnosis: organ with at least 13 lateral units helically arranged. Lateral units decreasing distally in size. Bracts elliptical to ovate in shape, up to 10 mm long. Ovules attached to the organ axis, up to 8 mm long, with a smooth outline and an asymmetrical base.
Description: The organ is cylindrical in form, 38 mm long and 16 mm wide, including more than 13 lateral units (Figure 1a). The lateral units are helically arranged along the organ axis, decreasing distally in size (Figure 1a). The lateral units diverge from the organ axis at angles between 27 • and 66 • (Figure 1a). Each lateral unit comprises a subtending bract and an axillary ovule (Figure 1b,c,g-j). The bract is 7 to 10 mm long, elliptical to ovate in shape, almost fully enclosing the ovule from bottom and laterals, leaving a ventral gaping suture (Figure 1a-j). Each ovule is attached to the organ axis by a short funiculus (Figure 1a-f,i,j). The ovule is anatropous, 7-8 mm long, and 1-2.3 mm thick, with a funiculus about 1 mm long and 0.3-1 mm in diameter (Figure 1b,c,g-j). The ovule is smooth-outlined and basally asymmetrical (Figure 1b,c,g-j; Figure 3).
Associated with the holotype of Combina is a part (Figure 2a,b) that resembles the lateral units in situ and thus provides an otherwise unavailable perspective on the lateral unit of Combina. The ovule and its funiculus in the holotype of Combina are not fully smooth and asymmetrical (Figure 1i,j): the funiculi appear to skew to one side of the ovule, suggestive of an anatropous ovule. This observation and inference of the holotype are further confirmed by an adaxial view of the isolated lateral unit (Figure 2a,b).

Discussion
The general morphology of Combina appears to be that of a coniferous cone, in which a lateral unit is composed of an axillary scale and a subtending bract. According to Schweitzer and Kirchner [19], there are only two Mesozoic genera with unilobate one-seeded lateral units, one is Drepanolepis and the other Ontheodendron. Since Ontheodendron has been recognized as a fossil stem [20,21], there is only one fossil genus left for comparison, Drepanolepis. Although both Combina and Drepanolepis appear similar to conifer cones in inverted ovules/seeds, Drepanolepis figured by Schweitzer and Kirchner is much more elongated and slenderer than Combina; most importantly, the bract encloses the axillary ovule to an unprecedented extent and forms an adaxial suture in Combina (Figure 1a,d,e), which is one of the implementations of the universal evolution trend of plant and organism reproduction [22]. Hitherto, the latter has not been seen in any known conifer cones (including Drepanolepis). This comparison justifies Combina as a new genus. Despite these differences between Combina and Drepanolepis, the comparison between them is meaningful for the homology of carpel and plant systematics (see Figure 4).
Currently, there are two competing hypotheses in botany on the homology and origin of angiosperm carpels, the Traditional Theory [4,6,23] and the Unifying Theory [24][25][26]. According to the former, a carpel is derived from a megasporophyll bearing ovules along its margins through longitudinal folding and enrolling [4,6,23]. According to the latter [24][25][26], a carpel is a composite organ comprising two parts of different nature, an axillary placenta (ovule-bearing branch) and a subtending ovarian wall (leaf). Although the Unifying Theory is in line with the results of function gene studies [27][28][29][30] as well as anatomical and morphological studies [16,17,31], it requires further independent observations (especially of fossil plants) for confirmation. Currently, there are two competing hypotheses in botany on the homology and origin of angiosperm carpels, the Traditional Theory [4,6,23] and the Unifying Theory [24][25][26]. According to the former, a carpel is derived from a megasporophyll bearing ovules along its margins through longitudinal folding and enrolling [4,6,23]. According to the latter [24][25][26], a carpel is a composite organ comprising two parts of different nature, an axillary placenta (ovule-bearing branch) and a subtending ovarian wall (leaf). Although the Unifying Theory is in line with the results of function gene studies [27][28][29][30] as well as anatomical and morphological studies [16,17,31], it requires further independent observations (especially of fossil plants) for confirmation.
It is evident that the ovules in Combina gen. nov. are borne directly on the organ axis (Figure 1b-j), not on the margins of any foliar part (bract), thus at odds with the widelyaccepted Traditional Theory, which expects ovules on the margins of a leaf. It is noteworthy that Combina is not the only evidence against the Traditional Theory, as (1) the carpels in the previously assumed ancestral angiosperm, Michelia (Magnoliaceae) and Illicum (Schisandraceae), have been shown to be composed of a foliar part and an axillary ovuliferous branch [17] or an ovule directly borne on the floral axis [16]; (2) in addition, another It is evident that the ovules in Combina gen. nov. are borne directly on the organ axis (Figure 1b-j), not on the margins of any foliar part (bract), thus at odds with the widely-accepted Traditional Theory, which expects ovules on the margins of a leaf. It is noteworthy that Combina is not the only evidence against the Traditional Theory, as (1) the carpels in the previously assumed ancestral angiosperm, Michelia (Magnoliaceae) and Illicum (Schisandraceae), have been shown to be composed of a foliar part and an axillary ovuliferous branch [17] or an ovule directly borne on the floral axis [16]; (2) in addition, another Triassic fossil reproductive organ putatively related to angiosperm, Nubilora, has its ovules directly borne on the floral axis [25]; (3) the carpel precursor assumed by the Traditional Theory (namely, megasporophyll) has not yet been found in either fossil or extant plants despite a century-long intensive painstaking search [32,33].
Several attempts have been made to raise hypotheses other than the traditional theory to account for the origin of carpels [3,24,25,31,34,35]. Regrettably, no plausible new morphological interpretation of carpel origin was given under the framework of the APG system, which was mainly based on molecular data. Although Sauquet et al. [1] discussed the arrangement of carpels in ancestral angiosperm flowers, they did not touch on the problem of the origin of carpels; their conclusion became controversial immediately and triggered criticism and heated debates among botanists [36][37][38]. Recent studies of extant basal angiosperms (e.g., Illicium [16] and Michelia [17]) reinforce that a carpel comprises axillary ovule(s) and a subtending foliar part, a conclusion that has long been suggested by gene function studies [29,30]. Integrating the outcomes of previous independent studies, Wang [24,25] proposed that a carpel is a composite organ derived through synorganizing an ovuliferous branch and a subtending leaf, a pattern that is workable for Amborella [25] (the currently assumed basalmost angiosperm [8,39]) as well as Magnolia (the traditionally assumed basalmost angiosperm [4,6]). The Late Triassic Nubilora [25] favors this generalization, as its ovules are attached directly to the cone axis and enwrapped by foliar parts. However, this single piece of fossil evidence appears weak against the overwhelming thinking inertia in botany. Now, with its axillary ovule directly attached to the organ axis and almost fully enclosed by its subtending bract, Combina gen. nov. seems to fit into the scenario depicted in figure 6.3e of Taylor and Kirchner [3] and the cases 23-25 in figure 8.40 of Wang [25]. The difference between Combina and the anticipated precursor of carpel [3] is restricted to its anatropous configuration of its ovule, which may be interpreted as a highly reduced and metamorphosed version of a former ovuliferous branch. It is noteworthy that a single axillary ovule is enclosed by a subtending leaf in Illicium from bottom and laterals [16], especially similar to our observation of Combina. Thus, Combina appears to be a piece of evidence supporting the Unifying Theory [25].
The most intriguing fact about Combina gen. nov. is that the bracts of Combina gen. nov. have started enclosing their axillary ovules, although not fully (Figure 1b-j). Therefore, although we are not sure whether the ovules of Combina gen. nov. are fully enclosed before pollination (if so, then Combina is a bona fide angiosperm) for the time being, Combina gen. nov. is apparently knocking on the door of angiosperms. It is noteworthy that Endress [40] recently took a carpel as the result of synorganization between a foliar part and ovule(s). This point of view is in full agreement with the implications given by Combina gen. nov. The possibility of the bract forming an outer integument in Combina can be easily excluded by the following facts: (1) the bract is directly attached to the organ axis rather than to the funiculus (ovule base) in Combina; (2) the bract and ovules have distinct and separated contours in Combina (Figure 2a,b). At the right time and with the right morphology, Combina gen. nov. appears to be an ideal precursor for angiosperm carpels, as it seems to have completed the evidence chain for carpel origination from a gymnospermous ancestor. The resemblance between Combina gen. nov. and the Palaeozoic conifers and Cordaitales [41]) (Figure 4a-l) seems to suggest that at least some angiosperms may share a common ancestor with some "conifers" and Cordaitales, if the previous axillary ovuliferous branches (as in Cordaitales [41], Palissya [42][43][44] , Metridiostrobus [45], and Stachyotaxus [44,46]) are reduced into a single ovule. Although we admit that relating conifers to angiosperms is at odds with most systematists (who cannot offer a plausible solution for carpel origin, however), our current proposal appears optimal, at least in terms of carpel homology, in the current academic context. Although we cannot determine, for the time being, that Combina is an angiosperm or an ancestral angiosperm, and the ovules in Combina are not fully enclosed as in angiosperms, it is noteworthy that, among all fossil plants, Combina demonstrates an unprecedented way of ovule-enclosing similar to that in some angiosperms (Illicium). This information is helpful for botanists trying to piece together the picture of plant evolution.