An enantiornithine specimen from Las Hoyas (Lower Cretaceous, Spain) offers new data on such matters (Fig. 1 and Fig. 2). Las Hoyas is a valuable Lower Cretaceous Konservat Lagerstätten that has provided important specimens of algae, plants and animals. The outcrop is composed entirely of Upper Barremian continental limestone deposits of the La Huerguina Formation. Freshwater, terrestrial and aerial organisms occur together in the same sedimentary horizon [17] and [18]. Las Hoyas has provided exceptional evidences of Mesozoic birds: initial report on Iberomesornis romeralli was followed by description of Concornis lacustris and Eoalulavis hoyasi [19]. These three aves have been recently placed phylogenetically within Enantiornithes, and Concornis and Eoalulavis within Euenantiornithes [3] and [19]. All three specimens are almost complete skeletons. Other isolated bony elements were retrieved at Las Hoyas, such as the isolated pedal skeleton attributed to I. romeralli [16] and [18], and LH21006 a-b, a slab and counterslab with an articulated tibiotarsus and tarsometatarsus. Both elements show the posterior view of a right hindlimb.

Herein we describe LH21006 a-b, a rather poorly preserved specimen with patchy bone preservation and impressions (see Fig. 1), in which some specific anatomical features can nevertheless be recognised. In the specimen the tarsometarsus (Tmt) is broken at its distal condylar area. The maximum length of Tmt is 22.9 mm. On the other hand, the tibiotarsus (Tit) has its proximal end broken and eroded, with a maximum preserved length of 23.5 mm. The skeleton of LH21006 recall Concornis lacustris in its proportions (Tmt = 22 mm and Tit = 35.36 mm), but these elements are smaller in Iberomesornis (Tmt = 11–14 mm; Tit = 20 mm). The equivalent size of LH21006 and Concornis suggests that the specimen is an adult individual of about 70 g body mass.

The specimen apparently has a true tibiotarsus, with proximal tarsals fused to the tibia. The tarsometatarsus is slender (straight and long) as in Concornis and the unfused metatarsals II and IV are situated in the same plane. They have sub-equal shaft diameters, being thinner than metatarsal III. The central portion of metatarsal III is rather convex and slightly more anterior to the other two. This new avian specimen shows an intercondylar sulcus (character #152 from [3]), an Ornithothoracine feature. The impression of a deep condylar area, divided by an intercondylar groove that is filled by sediment, can be observed on the specimen. No definite apomorphy shared with Enantiornithes can be demonstrated. However, the lack of fusion between metatarsals II to IV, a primitive character otherwise absent in Ornithuromorpha, suggests an enantiornithine affinity. Therefore, the specimen could be phylogenetically placed within Ornithothoraces, but not among Ornithuromorpha, which implies an enantiornithine assignment. This conclusion is strongly supported by the general similarity that the specimen has with Concornis lacustris, to which it is tentatively referred here. Given the fully ossified epiphyseal regions and the compact, smooth surface of the bone shafts, the material is presumed to belong to an adult individual.

After careful recording of the selected bone, including measurements, photographs and casting, slab and counter-slab were reunified in their original relative position in order to benefit from the total bone material available on both. The slabs were embedded in a polyester resin according to current practice (e.g., [21]) and allowed to polymerize slowly under controlled temperature. After hardening, excesses of resin and stone were sawed away from the block. The trimmed block was then subserially cut at slow speed with a circular diamond saw (Isomet Buehler LTD Company). Nineteen cross-sections were processed from the tarsometatarsal shafts: they offer a reasonably consistent picture of their structural pattern. In addition, tiny bone splinters from the same material were carefully collected, lined up and embedded together, to allow additional cross sections in more cortical material.

Casual observation would suggest a primary cortex entirely composed of an OCL, with clear evidence for at least two (and perhaps three) LAGs, and a free marrow cavity surrounded by a well-defined centripetal coating of endosteal bone (Fig. 3). This suggests a structural situation very similar to the one already described in larger Enantiornithes [4], [5] and [6], including the occurrence of LAGs. However, close observation reveals a more complex situation. The outer cortex is indeed formed of a rather typical parallel-fibered bone, with sparse flattened osteocytic lacunae and no vascularization (Fig. 4). This structure matches what is observed as forming the OCL in most extant small to tiny birds at somatic maturity [11]. It also agrees with earlier descriptions of presumably adult enantiornithine cortical bone [4] and [6]. However, the deeper cortex is rather different: it progressively contains a much more plentiful component of osteocyte lacunae with plump, rather than flat shapes (Fig. 4). This region is also permeated by a few vascular canals surrounded by bone lamellae, giving them an osteonal structure (Fig. 5). Relationships of the osteonal material to the neighbouring primary tissue strongly suggest erosion/reconstruction cycles: hence the osteons would be secondary (Fig. 5). In some sections, it is possible to follow extensions from the endosteal bone tissue into the deep cortex, where they form the secondary osteons (Fig. 6).

Some histology-independent characters already noted (see § Material) suggest an adult or nearly somatically adult condition for the specimen, in agreement with the histological structure of the OCL-like external cortex. Nevertheless, the inner cortex (also labelled as OCL on Figs. 3–6) still records a phase of the bird's ontogeny when radial bone apposition was more active. This phase is represented by the bone tissue vascularized by secondary osteons and much richer in plump-shaped bone cells than the external (slow-growing) cortex. Indeed, extensive comparative considerations among primary bone tissue types suggest this [1], [9], [13] and [14]. This structure represents evidence of the transition from the fast growth phase in hatchlings, as already documented by Chinsamy and Elzanowski [5], to the slow growth phase in adult Enantiornithes [4] and [6]. Because the structures still retained here in the deep cortex of a presumably sub-adult to adult specimen would record this intermediate phase in growth dynamics, their very presence also suggests that all the earlier phases of growth (not present here because of endosteal bone resorption followed by endosteal centripetal deposition) were also characterized by a fairly high rate, perhaps during a short time, during the juvenile stage of this enantiornithine. In connection with the above, it is noteworthy that Chinsamy et al. [4] and [6] already noticed and described two populations of osteocyte lacunae in their enantiornithine bones, namely a majority of flattened cell spaces and a few larger plump cell spaces. Apparently these two kinds of cell lacunae were not clearly set apart spatially in the cortex, but rather mixed up in their material. They would likely match the ‘plump’ and ‘flat’ cell lacunae that in our material are much better set apart spatially, in the deep and external cortex respectively, and that in our view suggest a progressive change in radial growth dynamics of the bone. Qualitatively, the number of cell spaces per volume unit of tissue in the deep and external parts of the cortex also concurs with this interpretation. The development of a rather extensive amount of finely fibered endosteal bone tissue around the marrow cavity and the occurrence of vascular canals organized as secondary osteons in the deep cortex of those small bones both concur with the anatomical data and occurrence of LAGs, suggesting at least a nearly adult somatic condition for the specimen.

Taken at face value, the data reported here suggest that in some small Enantiornithes a high initial growth rate could have been retained after hatching, allowing a significant percentage of total body size to be reached quickly (in a matter of days or weeks), while later growth could have been much slower and protracted, possibly during several years, taking into account the several LAGs and assuming them to be annual (which remains a debatable issue among birds [12]). Nevertheless, the non-vascularized OCL amounts to at least 50% of the cortical thickness, a high relative and absolute value for a bird [11]. It suggests (a) that growth speed could have decreased significantly when the bird had reached only half of its maximal body size (1/8 in body mass), and (b) that compact bone tissue thickness without vascularization could have been higher among Enantiornithes than among extant birds, an issue which deserves further inquiries. Although these conclusions are obviously tentative, they could nevertheless be tested if additional material of a morphologically more juvenile specimen from the same origin ever became available for histological analysis. Currently, very little is known about the growth dynamics of Enantiornithes. They did differ from extant birds but the few bone histological descriptions now available (and the interpretation of the adult bone cortex organization in light of the bird-specific bone growth dynamics and growth trajectories) do not necessarily suggest that Enantiornithes should have had a growth regime quite distinct from extant birds of commensurate body size–nor, by implication, completely different metabolic patterns or regimes, as previously suggested [2], [4], [6] and [7].