An encrusting kleptoparasite-host interaction from the early Cambrian

Parasite–host systems are pervasive in nature but are extremely difficult to convincingly identify in the fossil record. Here we report quantitative evidence of parasitism in the form of a unique, enduring life association between tube-dwelling organisms encrusted to densely clustered shells of a monospecific organophosphatic brachiopod assemblage from the lower Cambrian (Stage 4) of South China. Brachiopods with encrusting tubes have decreased biomass (indicating reduced fitness) compared to individuals without tubes. The encrusting tubes orient tightly in vectors matching the laminar feeding currents of the host, suggesting kleptoparasitism. With no convincing parasite–host interactions known from the Ediacaran, this widespread sessile association reveals intimate parasite–host animal systems arose in early Cambrian benthic communities and their emergence may have played a key role in driving the evolutionary and ecological innovations associated with the Cambrian radiation. Parasitic interactions are difficult to document in the fossil record. Here, Zhang et al. analyze a large population of a Cambrian brachiopod and show it was frequently encrusted by tubes aligned to its feeding currents and that encrustation was associated with reduced biomass, suggesting a fitness cost.

P arasitism is an enduring symbiotic relationship in which the parasite is nutritionally dependent upon the host for at least part of its life cycle, increasing its own fitness in the process and directly impinging upon the biological fitness of the host [1][2][3][4] . Parasite-host interactions form a significant proportion of the biotic interactions in extant global ecosystems, influencing many characteristics of species networks including behavior, population structure, and ecological function [5][6][7][8][9] . The antagonistic relationship between parasites and hosts has also been proposed as the primary mechanism leading to the evolution and maintenance of sexual reproduction, due to the negative frequency-dependent selection associated with parasitism 10 . Despite its obvious importance, the origins and early evolution of metazoan parasitism remains enigmatic 4,11,12 . Molecular phylogenies predict the emergence of parasitic clades in the Cambrian 13,14 and putative instances of shell damage, shell scarring and occasional bioclaustration from the early Cambrian 15,16 represent circumstantial evidence that hint at possible parasitism, but the rarity of well-preserved specimens precludes decisive identification of parasite-host interactions in the earliest Phanerozoic. Possible examples of epibiontism 17,18 , commensal infestation 19 and hitchhiking 20 are also known from the early Cambrian, but none of these constitute definitive instances of parasitism with a clear negative biological effect on the host. This absence of clear evidence for parasitism in the earliest animal communities may, in part, be due to a lack of cross-sectional quantitative analyses on Cambrian material of the type that have been demonstrated as necessary to identify and discriminate instances of animal parasitism in deep time 21,22 .
The early Cambrian (Stage 4) Guanshan Konservat-Lagerstätte occurs mostly in the lower 40 m of the Wulongqing Formation, which crops out over a geographically wide area in eastern Yunnan, located in southern China ( Supplementary Fig. 1). The Guanshan Biota is unusual in being proportionately dominated by brachiopods 23 , and so, is strongly differentiated from other Cambrian Konservat-Lagertätten such as the Chengjiang, Sirius Passet, Emu Bay Shale and the Burgess Shale, which are euarthropod-dominated assemblages [24][25][26][27] . The organophosphatic linguliform brachiopod Neobolus wulongqingensis sp. nov. is the most numerically abundant taxon in the Wulongqing Formation, with many thousands of specimens forming dense concentrations of monotypic, mostly conjoined shells, clustered closely on bedding plane surfaces ( Fig. 1a  seldom directly observed in the fossil record because taphonomic biases generally impede the preservation of direct interaction between organisms (see refs. 4,21,22,28,29 for exceptionally preserved cases in younger strata). The high-fidelity preservation and great abundance of specimens in the Wulongqing Formation ( Fig. 1 and Supplementary Figs. 2, 3, 5) provides a rare opportunity to investigate this unique interaction between a brachiopod host and their associated encrusting tube-dwelling organisms.
Here we assess differences in biomass between brachiopod individuals of the species N. wulongqingensis encrusted with tubes and those individuals lacking tubes, with biomass representing a proxy for the biological fitness of an individual. Our analyses suggest that the tube-dwelling organisms reduced the biological fitness of the host and, when considered in combination with observations of the preferred growth orientation of the encrusting tubes, these results suggest the interaction between the tubedwelling organisms and their host brachiopod represents kleptoparasitism. This instance in a Cambrian epibenthic marine community likely represents the oldest known parasite-host relationship in the fossil record and reveals that parasite-host interactions emerged in conjunction with the rise of the earliest animal communities during the Cambrian radiation.  Fig. 4f-h; see also ref. 23 ) in the brachiopods indicates rapid burial and minimal transport by episodic obrution deposits 18 . Despite this, the soft body of the tube-dwelling animal is not well-preserved, and its biological affinities are not self-evident. The greyish-white tubes, normally flattened by post depositional compaction, are immediately apparent (Fig. 1a- Fig. 3e-f), encrust the exterior of both dorsal and ventral valves of N. wulongqingensis ( Fig. 1f and Supplementary  Fig. 3d) with the open apertures exclusively oriented toward the anterior commissure of host brachiopods, indicating an intimate, life-long, in-vivo association. The tubes exclusively encrust the exterior of the host shell, which occasionally shows signs of minor damage or disruption of shell growth lines, but there is no evidence of boring into the interior of the brachiopod by the tubedwelling organism. The tubes are not found attached to any other hosts or substrates, such as the trilobite (Fig. 1g) or palaeoscolecid exuviae that occasionally occur in the shell beds. Consequently, we interpret this interaction as representing an obligate relationship, as defined by Poisot et al. 30 , as there is no evidence to suggest that the tube-dwelling organisms can adopt a free-living lifestyle in the absence of their brachiopod host.

Results and discussion
Impact on host biomass. Bayesian estimation analysis 31 demonstrates that a credible difference in biomass exists between brachiopods with encrusting tubes (n = 205) compared to those without (n = 224). There is no overlap in the 95% highest density interval (HDI) of the posterior distribution for the means of the two groups and the HDI for effect size does not overlap with zero ( Fig. 2a, b). Mean biomass for individuals with encrusting tubes is thus credibly lower than for those without tubes. A null hypothesis significance testing approach also identifies a significant difference between encrusted and non-encrusted individuals with a small effect size (Wilcoxon-Mann-Whitney W = 18,186, P = 0.0002, Cliff's Delta = −0.208). We therefore contend that individual brachiopods encrusted with tubes have reduced fitness when compared with their non-encrusted counterparts. On the basis of the difference in the values for mean biomass between the two groupings, encrustation results in a 26.08% reduction in overall fitness across the entire measured cohort.
Although our analyses indicate that brachiopods with encrusting tubes are reduced in biomass compared to those without, Log (attachment distance)  Fig. 6a). In some symbiotic relationships, the impact on the host is amplified depending on the number of parasites present 8 , but this relationship can be highly variable 32 . For our dataset, the biomass of the brachiopod host decreases when a single tube is encrusted to the shell surface, but no further decline is associated with an increasing number of tubes ( Supplementary Fig. 6a). Both proxies for increasing total parasite load also show no correlation with biomass (adjusted R 2 for both = −0.0003; Supplementary Fig. 6b, c). This suggests the tube-dwelling organism did not directly inhibit the feeding capability of the host, as larger numbers of parasites do not result in decreased fitness. However, a significant relationship exists between the attachment point of the encrusting tubes and the biomass of the brachiopod host (P = 2.2e−16, F = 327.6, 165 degrees of freedom, R 2 = 0.66; Fig. 2c), indicating encrustation earlier in ontogeny results in greater reduced biomass relative to hosts that have been infected at later ontogenetic stages, regardless of the number of symbionts present ( Supplementary  Fig. 6). In living brachiopods, smaller individuals generally display an increased growth rate compared to larger individuals 33 . It would therefore be expected that the impact on fitness would be greater for host individuals that are settled by parasites during earlier ontogenetic stages. The increase in median attachment distance for larger numbers of symbionts ( Supplementary Fig. 6d) and the larger size of specimens with greater than four tubes ( Supplementary Fig. 6a) establishes that higher infection rates can only occur when brachiopod hosts have already managed to grow to larger adult sizes and there is sufficient brachiopod shell surface area to accommodate a larger number of encrusting tubes.
This also indicates that the tube-dwelling organisms do not preferentially encrust smaller brachiopod individuals, as brachiopods are clearly encrusted in large numbers later in their ontogeny, when they have reached larger sizes. Our analyses demonstrate that the tube-dwelling organism directly impinges upon the biological fitness of the host, supporting the assertion that the encrusting tube-dwelling organisms are parasitic, rather than being either mutualistic or commensal with the brachiopod host. A reduction in host biomass or growth rate has been directly attributed to the presence of a parasite in a variety of extant symbiotic relationships 6,8,9 . Parasites typically increase the energetic requirements of infected organisms, as the host must generate sufficient energy to not only maintain its own requirements but also the needs of the parasite 6 . This commonly leads to hosts with decreased biomass when compared with uninfected individuals. This result represents the first definitive and statistically supported instance of parasitism from the Cambrian and indicates that parasite-host systems were well established by Cambrian Stage 4, suggesting this type of interaction probably emerged even earlier during the main pulse of the Cambrian radiation.
Variations in biomass between individuals and assemblages of the same species have also been previously attributed to regional variation in environmental stressors 6 . All specimens of N. wulongqingensis included in this analysis occur in dense aggregations (estimated 60,000 individuals per m 2 -Supplementary Fig. 5) from the same geographic locality ( Supplementary  Fig. 1a, locality 3) and stratigraphic package with similar sedimentological features 23 subject to similar environmental and depositional conditions. Consequently, the reduced biomass of tube-encrusted N. wulongqingensis individuals cannot be attributed to environmental factors and a parasitic affect is the most strongly supported probable cause.
A kleptoparasitic relationship. In all instances, the apertures of tubes are orientated toward the brachiopod commissure, spanning an arc (plan view) of~150° (Fig. 3). No tubes have been observed orientated toward the hinge line of the brachiopod. Tubes consistently grow beyond the commissural margin of N. wulongqingensis into, and slightly above but rarely beyond, the brachiopod chaetal fringe (Fig. 1b, c and Supplementary  Fig. 3a-c). Critically, the dominant growth direction of the tubes aligns tightly along a vector between 40°and 70°either side of the median plane of symmetry of the brachiopod (Fig. 3c; Supplementary Data 3); this alignment is most pronounced in shells with a single encrusting tube (Fig. 3d) but the same orientation pattern occurs in shells with all numbers of tubes ( Supplementary  Fig. 7), strongly supporting a preferential growth direction in the tubes toward the antero-lateral margin of the brachiopod shell. Five specimens of N. wulongqingensis from Wuding Quarry ( Supplementary Fig. 1a, Locality 2) preserve a partial spirolophe lophophore (Fig. 3a, b and Supplementary Fig. 4f-h). A spirolophe lophophore produces two separate inhalant laminar feeding currents at the antero-lateral edge of the shell margin 34 that match the preferred orientation and growth position of the encrusting tubes on shells of N. wulongqingensis (Fig. 3c, d and Supplementary Fig. 7). The preferred orientation of growth demonstrates that the tube-dwelling organisms were not purely utilizing the brachiopod as a hard substrate on which to construct their tubes. This data when combined with the demonstrated empirical cost to the host in the form of reduced biomass (Fig. 2a,  b), strongly supports kleptoparasitic behavior 35 . Kleptoparasitism is a form of competition, where food that is either already in the possession of the host or which the host has expended energy on obtaining and capture is imminent, is stolen by the parasite 35 . In our scenario, this involves the tube-dwelling organisms acting as intercept feeders-stealing a portion of the brachiopod feeding stream before it reached the chaetal fringe (Fig. 4). Iyengar 35 recognized six distinct morphological, behavioral and physiological criteria that characterize living sedentary/sessile kleptoparasitic interactions. At least five of these criteria can be directly applied to the relationship between the encrusting tube-dwelling organism and N. wulongqingensis (Table 1) further reinforcing a kleptoparasitic relationship.
Kleptoparasitism is rarely identified in the fossil record 22 , and no instances of kleptoparasitism, as far as we are aware, have been proposed for Cambrian communities 11,12 . Detailed empirical investigations of the energetic and nutritional cost of kleptoparasitism to the host, even for extant systems, are few 8,35 . For this reason, it is currently difficult to assess if the reduction in host fitness (~26%) we detect for N. wulongqingensis is typical of sessile kleptoparasitic relationships. Brachiopods are particularly vulnerable to exploitation by kleptoparasites, since active filter feeding represents the greatest energy expenditure in the life of brachiopods 36 , and the time lag between collection and ingestion of nutritionally beneficial particles also provides potential for other organisms to exploit this resource 35 . Combined with the fact that the biotic interaction we document is interpreted as obligate for the parasite, this suggests that the effect we observe is likely greater than would be the case in facultative kleptoparasitic associations. Intriguingly, obligate kleptoparasitism is exceedingly rare in modern marine systems 35 , which might suggest that this novel ecological relationship is always rare in benthic communities or has been secondarily lost some time during the Phanerozoic.
Verification of this kleptoparasitic relationship reveals that the heritage of parasite-host interactions can be traced back more than half a billion years to the rise of bilaterian animal communities during the Cambrian and further establishes the importance of the early Cambrian as a primary source of ecological novelty. Antagonistic biotic interactions have also been proposed as the drivers of widespread evolutionary phenomena such as the maintenance of sexual reproduction and genetic polymorphism at disease loci 10,37 . Both of these phenomena are known drivers of biodiversity increase, suggesting that the already established presence of parasitic relationships in Cambrian communities potentially had a fundamental role in the upsurge in evolutionary innovation associated with the Cambrian Radiation.  Fig. 4a) from the Gaoloufang section.

Methods
Studied material. High-density shell-bed concentrations of N. wulongqingensis and associated encrusted tube-dwelling organisms from the Guanshan Konservat-Lagerstätte were sampled as slabs of fine grained mudrocks and siltstones from thin-bedded mudrock layers within the basal 20 m of the Wulongqing Formation (formerly Wulongqing Member of Canglangpu Formation) from the Gaoloufang section, near Kunming ( Supplementary Fig. 1a). Previous intensive excavations of the Guanshan Konservat-Lagerstätte from the Gaoloufang and Kanfuqing sections near Kunming and Malong 23 have revealed a biota dominated by a rich assemblage of trilobites, bivalved arthropods, palaeoscolecid worms and an assortment of nonmineralized soft-bodied organisms. Aggregated high-density shell-bed concentrations of N. wulongqingensis have now been recovered at a number of localities ( Supplementary Fig. 1a) but all specimens included in this investigation come from rock slabs collected from the Gaoloufang section ( Supplementary Fig. 1a)  Measurements. Specimens were photographed using a Zeiss Smart Zoom 5 Stereomicrographic system and micro-XRF elemental mapping was undertaken using a Bruker M4 Tornado Micro-XRF spectrometer. The maximum linear dimension for each specimen was determined by taking measurements using ImageJ 1.49v 38 . Previous studies have shown that maximum linear dimension is directly correlated with soft tissue mass 39 , allowing direct conversion between linear measurements of shell size and biomass. We use measurements of biomass for individual specimens as a proxy for the relative fitness of brachiopod individuals and to identify differences in fitness between individuals with encrusted tubes and those without. We intended to combine these measurements with growth rate to obtain a holistic appraisal of the potential effects of encrusted tubes on brachiopod fitness, but the quality of preservation of N. wulongqingensis from the Guanshan Konservat-Lagerstätte is insufficient to consistently discern the very fine individual accretionary growth lines on individual specimens. We therefore confine our analyses to an assessment of potential disparities in biomass. To convert our linear measurements to biomass (specifically ash-free dry weight), we use the scaling coefficient and exponent for brachiopods presented in Payne et al. 40 . Body size measurements and biomass values for all specimens are provided in Supplementary Data 2.
Statistical analyses. All analyses of our dataset were undertaken using R 3.6.1 41 . To assess whether a difference in biomass exists between individuals with encrusted tubes compared to those without, Bayesian estimation analysis, using the 'BEST' package 31 generated complete distributions of credible values for group means and effect size (Fig. 2). In addition to this Bayesian approach, we also employ a null hypothesis significance testing approach, in the form of a Mann-Whitney U test, as a secondary assessment of the potential difference between our two groups. Cliff's Delta 42 , calculated using the 'effsize' package, is used as a measure of effect size for this secondary analysis.
Measurements of encrusted tube length, width, orientation and distance from margin were all compiled using the approach illustrated in Supplementary Fig. 8. Tube length and width were used to calculate total tube width per individual and the total surface area of each host shell covered by encrusted tubes; both values taken to represent proxies of total potential parasite load. Results of all these measurements are presented in Supplementary Data 2. Potential correlation between brachiopod biomass and total encrusted tube width per individual, total surface area of each host shell covered by encrusted tubes or attachment distance was assessed by fitting a simple linear regression model to the data ( Supplementary  Fig. 6b, c). Rose diagrams of encrusted tube orientation ( Fig. 3 and Supplementary  Fig. 7) were generated using the 'circular' package 43 .
Nomenclatural acts. This published work and the nomenclatural acts it contains have been registered in ZooBank, the proposed online registration system for the International Code of Zoological Nomenclature (ICZN). The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix "http://zoobank. org/". The LSID for this publication is: urn:lsid:zoobank.org:act:387EDCD0-3DA8-4965-8F36-A33032317209.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
The authors declare that all data supporting the findings of this study are available within the paper and its supplementary information files (specifically Supplementary Data 1 and  2). Specimens are deposited in the collections of the Early Life Institute and Department of Geology, Northwest University (prefix: ELI). Photographic material of the studied material is available from the corresponding authors upon request. The source data underlying Fig. 2 and Supplementary Fig. 6 is provided in Supplementary Data 1-3.

Extant sedentary kleptoparasitic interaction Guanshan brachiopod-encrusted tube interaction
Interactions are long in duration, the kleptoparasite utilizes few hosts within its lifetime (sometimes only one) Obligate and enduring interaction between kleptoparasitic encrusting tubedwelling organisms and adult organophosphatic brachiopod host Extreme negative impact on the host is not possible because of limited ability to find a new host if the present one dies Measurable negative non-fatal impact on host (Fig. 2)