Developmental dynamics is revealed in the early Cambrian arthropod Chuandianella ovata

Summary Segmentation and tagmatization have contributed to the preeminent success of arthropods since their first appearance in the Cambrian. However, the exact mechanism of segmentogenesis is still insufficiently known in living and extinct groups. Here, we describe the postembryonic development of a Waptiid arthropod Chuandianella ovata from the early Cambrian Chengjiang biota, South China. The new data illuminate a complex dynamic pattern of anamorphosis and epimorphosis, and a three-step process of segmentogenesis, i.e., the elongation of the terminal segment, delineation of an incipient segment, and full separation of a new segment. Compensatory growth is accomplished by rapid growth of new segments and/or generation of additional segments, which results in the trimorphism of the posterior tagma. Such complex developmental dynamics has rarely been known in the arthropod fossil record and its presence in early history helps to understand the rapid diversification of arthropods in the early Cambrian.

In living crustaceans, the mechanism of postembryonic segment formation has been permitted among the conchostracan Limnadia stanleyana and the anostracan Artemia salina, for which trunk segments are added progressively from the ''proliferative zone'' lying in front of the telson (Anderson, 1965(Anderson, , 1966. However, in the fossil record, most knowledge on arthropod development comes from the trilobites (Dai and Zhang, 2013;Dai et al., 2016Dai et al., , 2017Fusco et al., 2004Fusco et al., , 2012Hughes, 2007;Hughes et al., 2017) and phosphatized Orsten-type fossils (Haug et al., 2009(Haug et al., , 2010Mü ller and Walossek, 1988;Olesen, 2007;Stein et al., 2005Stein et al., , 2008Walossek, 1993;Walossek and Mü ller, 1998). Nevertheless, segment formation is rarely captured because of the gap of fossil materials. Except for Rehbachiella kinnekullensis (Walossek, 1993) from the Orsten fauna, few fossil taxa demonstrate the process of segment addition though ontogenetic sequences have been established, especially in many trilobites (Dai and Zhang, 2013;Dai et al., 2016Dai et al., , 2017Fusco et al., 2004Fusco et al., , 2012Hughes, 2007;Hughes et al., 2017;Shen et al., 2014;Zhang and Clarkson, 2009). Our investigation of Chuandianella ovata allows a reconstruction of its postembryonic development and recognition of the trimorphic posterior tagma varying from five to seven in the number of segments. Our material also provides the evidence for the segment genesis and compensatory growth of this species.

Intraspecific variation of the posterior tagma
The waptiid arthropod C. ovata is one of the iconic arthropods from the Chengjiang biota (ca. 518 million years before the present) . The anterior and middle tagmata are covered with a carapace, whereas the posterior tagma protrudes out the posterior end of the carapace (Figures S1 and S2) Shu, 2004, 2008;Hou et al., 2008Hou et al., , 2010Ou et al., 2020;Zhai et al., 2021). However, the middle tagma may be exposed when the carapace is dislocated from the body during the taphonomic process ( Figures 1C and  1D). There are 198 complete specimens showing the complete morphology of the posterior tagma varying from 4 to 7 in the number of segments. For the sake of description, we refer to the first segment in the posterior tagma as ''S 1'' and sort it gradually backwards.
Two small specimens show three limbless segments and a terminal segment with a pair of sub-elliptic terminal rami in the posterior tagma ( Figure 1). The length of each segment in the two specimens are measured as: S 1, 0.829 mm and 0.89 mm; S2, 1.055 mm and 0.997 mm; S3, 1.337 mm and 1.463 mm; S4, 1.623 mm and 0.771 mm (Table S1). The total length of the posterior tagma is approximately 3.7mm and 5.1 mm, and shorter than most specimens ( Figure 5A). Accordingly, they most probably represent immature specimens.
Fifteen specimens carry five segments in the posterior tagma, ranging from 2.898 to 13.055 mm in length (Figures 2 and 5A). In the posterior tagma, the first three segments become elongated toward the rear, whereas the length of the last two segments is variable in specimens. Four specimens are characterized by the S 3 being the longest and the length of S 4 and S 5 increasingly decreased (Table S1). In nine specimens the length of S 4 is the longest, significantly longer than S 5, and the length ratios of S 5 to S 4 (LS 5/LS 4) range from 0.48 to 0.95. Additional two specimens were featured by S 5 being the longest ( Figures 6A  and 6B), and therefore segments in the posterior tagma become increasingly elongated. The S 5/S 4 length ratio values are 1.20 and 1.29, respectively. Specimens with six segments in the posterior tagma are dominated in number. A total of 144 specimens have been recognized, accounting for 72% of specimens with a complete posterior tagma ( Figure 3). The length of the posterior tagma ranges from 2.82 to 12.229 mm ( Figure 5A). Similar to specimens with a five-segment posterior tagma, this largest group can be further subdivided into three categories according to the position of the longest segment. Two specimens with S 4 being the longest, in which the length of S 1 to S 4 gradually increases, whereas the length of the last two segments gradually decrease toward the rear. In these two specimens, S 5 is just slightly shorter than S 4, with the length ratio about 0.99. There are 138 specimens with S 1-5 becoming increasingly longer posteriorly. In these species, S 5 is the longest, and  iScience Article S 6 is consistently shorter than S 5. The length ratios of S 6/S 5 vary from 0.31 to 0.89. In addition, four specimens have a much longer S 6 ( Figures 6C and 6D), the length ratios of S 6/S 5 are 1.15, 1.32, 1.48, and 1.60, respectively (Table S1). Surprisingly, a superficial intersegmental groove is present in the longest terminal segment of ELI JS-027 ( Figures 6C and 6D), probably representing a prelude to separate a new segment.
There are 37 specimens with seven segments in the posterior tagma ( Figure 4), ranging in length from 2.379 to 10.608 mm ( Figure 5A). Two groups can be identified according to the length of segments. In 33 specimens, the length of S 1-6 is increasing gradually. S 6 is the longest, and S 7 is shorter than S 6. In the remaining four specimens, S 1-5 gradually increases in length, S 5 reaches a maximal length, and thereafter the last two segments decrease in length. There are two specimens with S 6 significantly shorter than S 5; the length ratios of LS 6/LS 5 are 0.39 and 0.41, respectively ( Figures 6E-6H). Therefore, these short S 6 likely represent newly generated segments.

Statistical analysis of the length of posterior tagma
To determine the differences among four categories of specimens with different numbers of segments in the posterior tagma, 198 specimens were used for principal component analysis (PCA). The result shows no variations among the four groups ( Figure S3B), indicating that they are identical in the length range of each segment in the posterior tagma. Among the four categories, the total length of the posterior tagma and each length of the last four segments are referred as independent (X) and dependent (Y) variables, respectively. The regression analysis shows: Line for the fourth segment to last: Y = 0.19296X + 0.060158, R 2 = 0.93246; Line for the antepenultimate segment: Y = 0.20724X + 0.077246, R 2 = 0.93559; Line for the penultimate segment: Y = 0.23228X + 0.10558, R 2 = 0.85075; Line for terminal segment: Y = 0.14821X + 0.092442, R 2 = 0.49731 ( Figure S3A). R 2 represents the coefficient of determination. The statistical analysis shows that the terminal segment has high dispersion, whereas the other three segments have good linear fitting, and the slope increases gradually from the reciprocal fourth to second segment. Thus, the length of the terminal segment has poor stability, and the rest three segments are gradually lengthening posteriorly. In addition, in the violin diagram ( Figure S3C) the length range of the terminal segment reveals a swelling ranging from 3 to 3.5mm in specimens with a five-segment posterior tagma, which are not present in the six-or seven-segment specimens. Hence, the length of the terminal segment has a wide range of variation, especially in specimens with a five-segment posterior tagma. Meanwhile, the vioplot also indicates that both the median and mean values of the penultimate segment are larger than those of the terminal segment among all four categories of specimens ( Figure S3C).

Anamorphosis and trimorphism
Morphological analyses reveal four phenotypes varying from four to seven in the segment number of the posterior tagma (Figures 1, 2, 3 and 4). The consistent morphological characteristics of the four phenotypes indicate that C. ovata does not possess metamorphic development, but follows an anamorphic iScience Article developmental mode because of segment addition during the postembryonic ontogeny. The phenotype with a four-segment posterior tagma is small, represented by only two specimens, whereas the sizes of these two specimens are both larger than the smallest individual among phenotypes with five-segments (Figures 1 and 5A). Hence, these four-segment specimens can be interpreted as immaturities of the fivesegment phenotype. Each of the other three phenotypes are represented by numerous specimens (Table  S1). Such interspecific variation in the number of segments were usually considered as a series of growth stages of a single phenotype, e.g., Fuxianhuia protensa (Fu et al., 2018), Rehbachiella kinnekullensis (Walossek, 1993), and many trilobites (Dai and Zhang, 2013;Dai et al., 2016Dai et al., , 2017Fusco et al., 2004Fusco et al., , 2012. However, this is less likely the case of C. ovata. First, measurements demonstrate that the posterior tagmata of these five-segment, six-segment, and seven-segment phenotypes have a similar size range from immaturities to adults, approximately 3 to 13 mm long ( Figure 5A). Moreover, principal component analysis indicated that the four categories of specimens could not be distinguished according to the length of each segment in the posterior segment ( Figure S4B). Third, the largest specimen (ELI SJZ-B23) carries a fivesegment posterior tagma ( Figures 2E and 2F), whereas the smallest (ELI SJZ-B16-604) bears a sevensegment posterior tagma ( Figure 5A; Table S1). The consistent morphology of the four phenotypes indicates that these specimens acquired adult morphology during the postembryonic development. Therefore, it is less likely that the significantly increased five-segment specimens represented ''giant larval'' as seen in the extant and extinct crustacean larvae (Gundi et al., 2020;Nagler et al., 2017). The size decrease iScience Article during the postembryonic development is presently unknown among extant and extinct arthropods. In addition, if the three phenotypes represent a successive developmental sequence, the postembryonic development of C. ovata would consist of three phases: the anamorphosis with the segment number of the posterior tagma increasing to five, followed by an epimorphosis during which the size increases whereas the segment number of the posterior tagma remains as five, and then anamorphosis again with addition of two more segments ( Figure 5C). Such a three-phased pattern of postembryonic development has not been reported in either fossil or living arthropods. Therefore, it is most probably that C. ovata has a trimorphic posterior tagma varying from five to seven in the segment number. The postembryonic development of each phenotype involves two phases: anamorphosis with both segment number and size increasing, followed by epimorphosis solely increasing size ( Figure 5B and 7).

Genesis of new segments
In our collection, two seven-segment (ELI JS-047 and ELI SJZ-B20-900A) and one six-segment (ELI JS-027) specimens demonstrate the genesis of new segments in the posterior tagma. In the two seven-segment specimens, the length of S 6 is significantly shorter than S 5 ( Figures 6E-6H), which contradicts the general trend of posterior segments elongating toward the rear. Moreover, both regression analysis and vioplot show a broader range in the length of the terminal segment, which are probably attributed to subterminal separation of a new segment. Consequently, the terminal segment was relatively longer upon separation and shorter immediately after separation ( Figure S3A and C). In our research, there are six specimens with a much longer terminal segment in the five-segment and six-segment phenotypes ( Figures 6A-6D; Table S1). Among them, the terminal segment of the specimen ELI JS-027 shows a shallow groove at the mid length ( Figures 6C and 6D), reminiscent of the case in the ''Orsten'' type crustacean R. kinnekullensis, for which the formation of trunk segment is characterized by two steps. Before a new segment finally delineates from the terminal end, there is an intermoult stage with fissure on the dorsal surface (Walossek, 1993). Therefore, it is reasonable to assume that posterior segments of C. ovata could be generated through three steps, including the elongation of the terminal segment, the formation of blurred incipient segment (shallow groove) and a separation of new segment in the subterminal region (Figure 7). Such segmentogenesis is basically concordant with living crustaceans Limnadia stanleyana and Artemia salina, for which postmandibular segments are progressively proliferated from a growth zone lying in the anterior region of the telson (terminal segment) (Anderson, 1966). The post-protaspid ontogeny and the pattern of segment release from the pygidium have been extensively studied in trilobites, such as Duyunaspis duyunensis (Dai et al., 2017), Eoredlichia intermediate (Dai et al., 2016), and Aulacopleura koninckii (Fusco et al., , 2012Hughes, 2007;Hughes et al., 2017). However, the exact mechanism of new segment genesis is still poorly understood. The present study provides valuable evidence for the postembryonic segmentogenesis in an early Cambrian arthropod.

Compensatory growth
Target phenotype refers to a phenotype determined by the individual genetic makeup without perturbing factors Nijhout and Davidowitz, 2003). Therefore, during the postembryonic development of arthropods, there should be a series of target character states at each stage of ontogeny, if external conditions such as temperature, nutrition, and parasitism are consistent . However, the majority of arthropods deviate from the target trajectory as a result of variations in the external factors Hartnoll, 1982;Nijhout and Davidowitz, 2003), and hence need to adjust via compensatory growth which is accomplished by altering the number of stages and/or a stage-by-stage feedback mechanism to reach the target features .
In our collection of C. ovata, more than 91% of the specimens have the longest penultimate segment (Table  S1), but the length of newly formed segments is significantly shortened (Figures 6E-6H). The same phenomenon was observed in growth stages of R. kinnekullensis, where the incipient segment was shorter (in Walossek, 1993, plate 14;17;20;23) (Walossek, 1993). Therefore, compensatory growth is necessary for newly separated short segments to achieve the target length during anamorphic phase. Compensatory growth is also the mainspring of the trimorphism seen in C. ovata. Statistic measurements demonstrate that the six-segment posterior tagma is the dominant phenotype ( Figure 5A), accounting for more than 72% of specimens. Accordingly, in the absence of compensation, only the six-segment posterior tagma would be present in the epimorphic phase. However, three phenotypes have been recognized in the adulthood of C. ovata ( Figure 5A). Specimens with a five-segment posterior tagma started their epimorphic development ahead of schedule because of the excessively rapid elongation rate of new segments, and hence reached their target size one stage earlier ( Figure 5B). Although specimens with a seven-segment posterior ll OPEN ACCESS iScience 25, 103591, January 21, 2022 7 iScience Article tagma entered the epimorphic phase one stage later, they grew to the target size through a slower elongation rate of new segments or relative rapid generation of new segments ( Figures 5B; S4C). Consequently, these compensatory growth modes are responsible for the trimorphic posterior tagma reaching the target size. Similar compensatory growth was reported in many extant arthropods (Hartnoll and Dalley, 1981;Klingenberg, 1996;Tanaka, 1981;Twombly and Tisch, 2000;West and Costlow, 1987). Typically, in the postembryonic development of the tobacco hawkmoth Manduca sexta and the cockroach Blattella germanica, individuals reach the target threshold size by increasing the number of developmental stages (Tanaka, 1981;Nijhout, 1975). The dynamics of segment generation and growth seen in C. ovata is as complex as extant representatives. The waptiid arthropod Chuandianella ovata from the 518 million years old Chengjiang biota carries a trimorphic posterior tagma with five to seven segments and demonstrates a complex developmental pattern. Each phenotype underwent an anamorphic phase during which a fixed number of segments in the posterior tagma are obtained, followed by an epimorphic phase without new segment addition when molting (Figure 7). The three phenotypes are comparable in size range, and the one with six segments is dominant in the number of specimens ( Figure 5A). C. ovata demonstrates a similar compensatory growth mechanism to living arthropods. In particular, during the anamorphic phase, the generation and growth of new segments were coupled dynamically, and hence C. ovata reached its target size through increasing or decreasing developmental stages ( Figure 5B). As a result, some had a segment growth rate exceeding its generation rate and thus started their epimorphosis relatively earlier at the stage with five segments in the posterior tagma, the majority went the epimorphic stage when the posterior tagma had six segments, whereas others had a relatively later epimorphic stage with one more segment when the segment generation rate exceeds its growth rate. Accordingly, the adults maintain the trimorphism in the posterior tagma (Figure 7). The newly generated segment was short, separated from the subterminal region of the terminal segment, and reached its normal length through an accelerating growth rate. The trimorphism and developmental pattern seen in C. ovata are rarely known in fossil records and reveals that the complicated dynamics of segment growth and generation evolved early in arthropod history.

Limitations of the study
Fossils used in our study are unlikely to represent a single population in the strict sense, which is always the case in paleontological studies. However, in our study, all specimens were collected from a narrow iScience Article stratigraphical range of the Eoredlichia-Wutingaspis trilobite biozone Zhang et al., 2007) and are generally considered to reflect the size and morphological range of ''a single population''.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

ACKNOWLEDGMENTS
We are grateful to T. Dai for discussion and C. Aria for assistance in Figure S3 analyses. Two anonymous reviewers are acknowledged for constructive comments. This work was supported by National Natural Sci-

DECLARATION OF INTERESTS
The authors declare no competing interests.