Keeping with no piece left intact at all — a sponge taken apart into single cells, its shape, its facing and its plumbing all erased, whose scattered cells still recognise their own, gather, sort, and — in the few cases that finish — become a filtering body again
R948 closed with a fragment that still held something: a piece of tissue, a map in its muscle, a facing at each wound. This page takes the turn it pointed to. Take a sponge — in the founding work, the marine demosponge Clathria prolifera, long called Microciona prolifera — and press it through fine silk mesh until there is no piece at all, only a suspension of single cells in seawater. Nothing in the dish has a head end, a surface, or a canal. What is kept, if anything is, has to be kept cell by cell.
First, what the sieve leaves. H. V. Wilson described the procedure and its sequel in 1907: tissue forced through bolting cloth dissociates, and the freed cells round up. A suspension looks poorer than the animal it came from — across studies only a handful of forms stay distinct, and the two found in every species are amoeboid cells of the archaeocyte kind and choanocytes, the collared, flagellated feeding cells, which at first keep collar and flagellum. Much of the rest has, in effect, taken off its uniform: many cells lose their specialised look soon after dissociation and take on the amoeboid form. The crowd in the dish is therefore not a labelled kit of parts. It is mostly wanderers, plus the feeders, and whatever each cell still carries on its surface.
Second, the cells know their own — and say so by touch. Given a substrate, the amoeboid cells put out pseudopodia and move; contacts accumulate, small clumps take in free cells and merge with other clumps. The recognition is specific. In Wilson’s mixed dishes, cells from differently coloured sponge species did not settle into permanently mixed animals: the stable masses that formed were preferentially of one species, each kind gathering with its own. Later work gave that preference a carrier: an enormous extracellular proteoglycan complex, the aggregation factor — in Microciona on the order of 2×107 in molecular weight — released into solution at dissociation and then bridging cell to cell through surface receptors. Part of its grip is carbohydrate recognising carbohydrate, in a calcium-dependent self-association shown directly for its sulfated disaccharide epitope; remove calcium’s specific place in that step and the association fails. The factor’s genes are highly polymorphic, which is part of why the system can act as a self / non-self test and not merely as glue. One caution belongs here: the factor is characterised in detail in a few demosponges, and its exact make-up varies across sponge groups — recognition is general, this molecule’s biography is not.
Third, a crowd becomes a ball before it becomes a body. The first products are primary aggregates: loose, irregular handfuls with much the same mixed contents as the suspension — cells, spicule fragments, fibres, symbionts carried over — and, pointedly, no building yet. Building starts at the surface. As an aggregate turns into a primmorph, its outer cells flatten and change trade into exopinacocytes, sealing the ball under a continuous outer skin, the exopinacoderm; in demosponges each surface cell makes that change separately and then joins its neighbours, and choanocytes recruited to the surface shed collar and flagellum to do it. Inside, the packing tightens. A primmorph is a real achievement — an inside and an outside, restored — and it is not yet a sponge: it has no working water system and does not feed as one.
Fourth, gathering is not enough — the crowd needs the right members, in enough number. When the freshwater sponge Ephydatia fluviatilis was separated into cell fractions, only the fraction containing archaeocytes went on to rebuild an intact sponge. A mixed fraction lacking archaeocytes stalled at the primmorph stage; a choanocyte-enriched fraction stalled earlier, at loose aggregates; a fraction lacking both did not reaggregate at all — a pattern also reported for the marine demosponge Hymeniacidon perleve. Size gates the outcome too, on reported figures that differ by species and study: on the order of a thousand cells in E. fluviatilis and two thousand in Clathria prolifera as cited minima for progress, and in one detailed E. fluviatilis study only aggregates wider than about 200 µm developed further, while smaller ones stalled or managed only a skin and a few chambers, late. And archaeocytes are not the whole story of rebuilding: choanocytes are a second source of plastic cells, differentiated cells of several kinds can change trade outright, and which source matters most varies by lineage — “one stem cell explains it” is more than any of these experiments shows.
Last, the finish is rare, and it is plumbing. Where primmorphs do go further, they build the aquiferous system: cells draw together into rosettes that hollow into choanocyte chambers, canals open as cells draw apart — or, in dense aggregates of species such as Halisarca dujardinii and Halichondria panicea, as interior cells die in a controlled way and leave the channel behind — and at last pores and an oscular tube appear and the ball begins actively filtering. What stands there is a small functional sponge, simpler than the parent, in the juvenile rhagon or olynthus grade of organisation, its axis and polarity made new rather than recovered: reaggregation always passes through that loss. The honest census is the close of the 2021 review: of more than forty species studied, only eleven are documented rebuilding a functional sponge this way; in one three-species comparison the early stages were shared and only H. dujardinii finished, the others arresting as early primmorphs or primmorphs. That is the contrast with R948 stated plainly: the planarian fragment kept the pattern distributed in tissue it never lost; the sponge keeps it, when it keeps it at all, as recognition carried on cell surfaces and a rebuilding programme its cells can re-enter — and often the crowd gathers, seals itself, and goes no further.
A walk from a suspension to a sponge. Five stages: a body erased to single rounded cells; recognition, species-preferring, carried by the aggregation factor’s calcium-dependent carbohydrate self-binding; loose primary aggregates sealing into primmorphs under a new outer skin; the gates — archaeocytes present, the aggregate large enough — that decide whether development proceeds; and, in the minority that finish, chambers, canals, pores and an osculum, filtering again. Schematic, compressed from the published accounts.
Walk it in order.
Schematic only: the stages compress work across sponge species, dissociation methods and laboratory conditions, so no single dish runs this exact account. Labelled cautions: do not say every sponge rebuilds from dissociated cells — most studied species arrest at aggregates or primmorphs, and only eleven of more than forty are documented finishing. Do not call a primmorph a functional sponge — it has a skin and an inside, but no working aquiferous system. Do not present the aggregation factor as one universal molecule — it is characterised in detail in a few demosponges and varies across groups. Do not present mixed-species sorting as instant or absolute — wrong-species contacts happen; it is the stable organisation that is species-preferring. Do not make archaeocytes the sole stem cell — choanocytes and transdifferentiating cells also build, and the balance varies by lineage. And do not treat the cell-count and diameter thresholds as constants — they are reported figures for particular species and studies.