Parental food provisioning is often indispensable to an offspring’s growth, development, and survival (
1,
2) and subsequently affects litter size, sex ratio, and offspring body size (
3,
4). Foods provided by parents can be classified into three forms: direct food sources, regurgitation feedings, and specialized food sources (
4). Specialized foods include mammalian milk, as well as trophic eggs in amphibians and invertebrates (
4–
6), epidermal mucus in fish (
7), and milk in cockroaches (
8) and pigeons (
9). However, these other foods are very different from mammalian milk in both duration of provisioning and degree of parent-offspring interaction. Parental food provisioning often stops when offspring acquire foraging ability (
4), and provisioning for nutritionally independent offspring has mainly been reported from longer-lived vertebrates (
4,
10). Extended food provisioning frequently co-occurs with prolonged parental care, enhancing offspring fitness by providing assistance in foraging (
11), teaching skills of foraging and competition (
12), or influencing the morphology and behaviors of offspring (
13). In extreme cases, parental care is extended to sexually mature offspring (
4,
10), but this has only been recorded in long-lived social vertebrates (
4). Prolonged care increases offspring fitness by enabling them to allocate more time for learning foraging and social skills (
14), antipredator behaviors, and defense against nest parasites and for selecting mates (
14,
15).
Toxeus magnus (Araneae: Salticidae) (
16) is a jumping spider that mimics ants. The breeding nest is composed of either several large individuals (fig. S1), with two or more adults, or one adult female and several juveniles. This is a puzzling observation for a species assumed to be noncolonial and suggests that
T. magnus might provide either prolonged maternal care or delayed dispersal. We thus questioned (i) whether
T. magnus evolved long-lasting maternal care; (ii) if it has, whether the mother provides food for the offspring; (iii) what benefits could be gained from maternal care; and (iv) how long the mother provides an effective maternal care.
We first assessed how offspring developed and behaved under maternal care in laboratory conditions. No spiderlings were observed leaving the nest for foraging until they were 20.9 ± 1.2 days (
N = 207,
Nnest = 19; SEM) old, and the mother was never seen to bring food back to the nest during this period, but offspring body lengths increased continuously (from 0.9 mm at hatching to 3.5 mm at 20 days,
N = 187). Closer observations revealed that the mother provided a seemingly nutritive fluid, hereafter called milk, to the offspring. During the first week, she deposited milk droplets, excreted from her epigastric furrow, on the nest’s internal surface. In response, spiderlings came and sucked the droplets (
Fig. 1B). After the first week, milk deposition stopped and the offspring sucked directly from the mother’s epigastric furrow (
Fig. 1C and movies S1 and S2). Although spiderlings sometimes left the nest to forage from about 20 days, milk sucking continued until they reached the subadult stage at 37.5 ± 1.3 days (
N = 167) (
Figs. 1C), with body length of 5.26 ± 0.06 mm (around 80% of adult body length). From 40 days, milk feeding stopped, but spiderlings still used the nest at night after reaching sexual maturity within the next 20 days. Adulthood was reached at 52.2 ± 0.6 days with a body length of 6.57 ± 0.06 mm (
N = 145) (
Fig. 1D and fig. S2). Milk provisioning in
T. magnus involves a specialized organ over an extended period, differing markedly from cockroaches, which deposit nourishing substances to the brood sac of developing embryos (
8). In these aspects,
T. magnus milk provisioning is more similar to mammalian lactation.
Observations under the microscope showed droplets leaking from the mother’s epigastric furrow where the spiderlings sucked milk (
Fig. 2, A and B). Spider milk total sugar content was 2.0 mg/ml, total fat 5.3 mg/ml, and total protein 123.9 mg/ml, with the protein content around four times that of cow’s milk (
17). We suggest this milk might have evolved from trophic eggs, unviable eggs functioning as a food for newly emerged offspring (
18). The epigastric furrow is the egg-laying opening of spiders (
6), and trophic eggs have evolved in diverse invertebrate taxa, including spiders (
6). Trophic eggs are usually a one-time investment (
4,
6), but mothers may also supply progressive trophic eggs, as in the burrower bug
Adomerus triguttulus (
18). However, progressive provisioning usually stops after hatching (
19), whereas spider milk is supplied continuously from hatching to subadult.
The mother continued nest maintenance throughout, carrying out spiderlings’ exuviae and repairing nest damage. When receiving both maternal care and milk, 76% of the hatched offspring survived to adulthood. We also recorded a highly female-biased adult sex ratio: 84.14% (122/145). Although the mother apparently treated all juveniles the same, only daughters were allowed to return to the breeding nest after sexual maturity. Adult sons were attacked if they tried to return. This may reduce inbreeding depression, which is considered to be a major selective agent for the evolution of mating systems (
20–
22).
When we blocked the mother’s epigastric furrow immediately with correction fluid after hatchling emergence, the hatchlings stopped development and died 10.4 ± 0.5 days later (
Fig. 3A and fig. S3), showing their complete dependence on the milk supply. A separate experiment showed that correction fluid did not affect spiderlings’ survivorship (fig. S6). We then tested why parental care and milk provisioning were continued after 20 days when the spiderlings were able to forage for themselves. We assessed the effects on survival, growth, foraging, and nest departure of (i) blocking the milk at this stage but leaving the mother in the nest to assess the effect of maternal care without milk provisioning and (ii) removing the mother from the nest to assess the combined effects of maternal care and milk provisioning. Removing the mother after 20 days significantly reduced survivorship (
Fig. 3A) and body size (
Fig. 3B) and increased foraging (
Fig. 3C) and activities outside the nest (
Fig. 3D). Blocking milk significantly reduced survivorship (
Fig. 3A), but did not affect body size (
Fig. 3B), and increased foraging (
Fig. 3C) and activities outside the nest (
Fig. 3D) compared with those receiving full maternal care and milk. Among spiderlings not receiving milk, those receiving maternal care survived significantly better (
Fig. 3A), grew faster (
Fig. 3B), and left the nest more rarely (
Fig. 3D) than those without the mother, but both showed similar foraging frequency (
Fig. 3C) (table S1 to S4). The milk-blocked spiderlings still attempted to suck milk from the mother (movie S3). These results show that from days 20 to 40, milk provisioning is not essential for spiderlings’ survival, as they can compensate by foraging more, but has a positive effect on survivorship and development in lab conditions. In the field, milk provisioning, by reducing time spent outside the nest, might also reduce predation risk, as smaller offspring are more prone to predation (
23). When spiderlings did not have milk, the higher surviorship with their mother present might be mediated by parasites. Although not quantified, parasites were observed on juveniles commonly without maternal care but rarely when the mother was present. Nest cleaning and repairing by the mother may partly explain this reduction in parasite loads. Nest parasites are common in invertebrates (
24,
25), and parental care usually plays a crucial role in reducing juvenile parasite infections (
4,
10,
25).
Removing the mother at day 20 reduced the percentage of spiderlings reaching maturity [normal maternal care (MC) = 75.7 ± 4.6%; mother removed at 20 days (MR20) = 50.4 ± 5.3%] and reduced their adult body size (by 12%), whereas blocking milk did not change their adult survivorship [milk blocked at 20 days (MB20) = 73.8 ± 6.4%] or body size (MB20: 6.64 ± 0.07 mm) compared with the control (
Fig. 4, A and D). Most notably, removing the mother greatly affected the adult sex ratio (female/total: MC = 0.84, MR20 = 0.50), whereas blocking milk did not (MB20 = 0.87) (
Fig. 4C). Further analysis showed that the reduction of offspring body size after mother removal is not caused by a sex ratio change coupled with intersexual differences (table S6). No treatment impacted the timing of developmental stages (MR20: 52.5 ± 0.78 days, MB20: 50.5 ± 0.82 days, MC: 52.3 ± 0.60 days) (
Fig. 4B and table S5). Thus, milk provisioning after 20 days does not affect adult survivorship, body size, sex ratio, and development time, but the mother’s presence plays a key role in assuring high adult survival and normal body size. This may reflect the parasite defense discussed above. The mother’s presence is also required for achieving the female-biased sex ratio of adult offspring. Sex role divergence theory suggests that the reproductive value of a population is heavily determined by the sex that invests more in parental care (
26–
28). In
T. magnus, the mother invests much more than the male invests, predicting a female-biased sex ratio to be optimal for reproductive success with a polygamous mating system. Unequal adult sex ratios could be passively induced, as with the selective mortality with respect to sex in the bark beetle
Dendroctonus ponderosae (
29), or maternally induced, as in the spider
Oedothorax gibbosus (
30), but the mechanisms remain largely unclear.
In mammals, lactation is likely an adaptation mitigating an uncertain access to food resource as a juvenile and compensating for low probabilities for future reproduction (
31). More generally, food provisioning helps promote offspring fitness by reducing risks linked to self-feeding (
32). In nonmammals, we suggest that the most important ecological conditions favoring the evolution of lactation might be predation risk and uncertain food access. As a response, the mother’s physiology, behavior, and cognition might have changed to adapt to providing milk and prolonged maternal care as in mammals (
33). Extended parental care could have evolved in invertebrates as a response to complex and harsh living environments that require offspring skills (e.g., hunting, predator defense) to be fully developed before complete independence. Another aspect to investigate is whether lactation and extended parental care are accompanied by a reduction in offspring number because of milk production and parental care costs. Lastly, we anticipate that the discoveries presented in this study will encourage a reevaluation of the evolution of lactation and its occurrence across the animal kingdom.
Acknowledgments
We thank an anonymous referee and A. Jordan, L. Li, G. Liu, X. Yao, U. Gelin, X. Wang, and J. Chen for suggestions, X. Yang for equipment assistance, the Central Laboratory and Horticulture Department at XTBG for spider collection and maintenance, and Z. Wang and P. Wen for spider milk analysis. Funding: Work was supported by NSFC (no. 31370452) and CAS 135 and CAS-SEABRI (no. 2017 XTBG-F03). Author contributions: Z.C., R.C.Q., and X.J. designed experiments. Z.C. collected data and wrote the original manuscript. R.T.C., T.C.D., and R.C.Q. revised manuscripts and figures. S.J.L. analyzed data. S.Z. and H.L. assisted experiments. R.L. and C.L. analyzed milk composition. Competing interests: We declare no competing interests. Data and materials availability: All data are available in the main text or the supplementary materials.