The larger catalogue for the whole series is on the parent page, the origin of life: the literature.
Sources cited
| Source | What it is | Access | Direction |
|---|---|---|---|
| Deamer 2002 | The chemistry of self-assembly, and the early statement that divalent cations block it, written conditionally. | Opened in full | complicates |
| Hanczyc, Fujikawa & Szostak 2003, Science 302:618 | Clay accelerates vesicle formation, is encapsulated, and carries adsorbed RNA inside; the vesicles grow and divide by extrusion. | Opened in full | undercuts |
| Chen, Salehi-Ashtiani & Szostak 2005, JACS 127:13213 | The earliest clear statement of the salt problem: aggregation at 0.2 M NaCl, precipitation below 1.5 mM magnesium. | Opened in full | supports |
| Zhu & Szostak 2009, JACS 131:5705 | Growth into threads and division under 0.35 dynes/cm² shear, contents retained, three cycles. | Opened in full | undercuts |
| Adamala & Szostak 2013, Science 342:1098 | The 50-200 mM magnesium figure and the flat statement that it destroys vesicles. Citrate as the fix. | Opened in full | supports |
| Deamer & Damer 2017, Astrobiology 17:834 | "Self-assembly of membranes… has not been demonstrated in hydrothermal vents." | Opened in full (CC BY) | supports |
| Milshteyn, Damer, Havig & Deamer 2018, Life 8:11 | The experiment behind that claim: vesicles in Yellowstone spring water, none in seawater. | Opened in full (CC BY) | supports |
| Joyce & Szostak 2018, CSH Perspect Biol 10:a034801 | The ≥4 mM instability figure, the citrate concession, and the summary that no protocell replicates its contents across generations. | Opened in full | supports |
| Cornell et al. 2019, PNAS 116:17239 | Serine and glycine protect decanoic acid vesicles at 10 mM magnesium, by driving them multilamellar. | Opened in full | complicates |
| Jordan et al. 2019, Nat Ecol Evol 3:1705 | Twelve-component amphiphile mixtures give vesicles at seawater ionic strength, 50 mM magnesium, 70 °C, pH 11-12. | Accepted manuscript opened in full | undercuts |
| Toparlak et al. 2023, ACS Nano 17:23772 | Cyclophospholipid protocells that grow, divide and extend a primer. Opens with the field's own diagnosis. | Opened (CC BY-NC) | complicates |
| Agrawal et al. 2024, Sci Adv 10:eadn9657 | Coacervate exchange times, and the rain-crosslinking fix that holds 35-mers for days. | Opened in full (CC BY) | complicates |
| Cohen et al. 2024, PNAS Nexus 3:pgae084 | Soda lakes cap magnesium near 1 mM; RNA polymerisation runs at comparable rates in lake water. | Opened (CC BY) | undercuts |
| Lee, Pir Cakmak, Booth & Keating 2024, Small 20:e2406671 | Coacervate-supported membranes intact at 25 and 50 mM magnesium. The strongest published answer to the magnesium argument. | Opened in full | undercuts |
| Joshi, Lee, Chan & Keating 2026, Langmuir 42:13582 | Heterogeneous amphiphile mixtures outperform pure ones in hybrid protocells. Does not test magnesium. Post-cutoff. | Opened in full (CC BY) | complicates |
Considered and not used
- Chen, Roberts & Szostak 2004, Science 305:1474, the vesicle-competition paper. The result is described in the article; the citation sits inside the Zhu and Szostak footnote rather than standing alone.
- Mansy et al. 2008, on membrane permeability to nucleotides. Described in the parent article; not needed for the argument here.
- Budin & Szostak 2011, on the phospholipid transition driven by selection. A striking result about how membranes could improve, and not about whether they form.
- Jia, Hentrich & Szostak 2014, OLEB 44:1, the coacervate exchange-time measurement. Not opened. Its figures reach this article through Agrawal et al. 2024, which was opened, and the footnote says so.
- Sacerdote & Szostak 2010, on phosphate permeability. Not needed.
- The meteoritic fatty-acid abundance literature. The chain-length arithmetic in the article is this project's own calculation on published abundance figures, and it is flagged as such in the text rather than footnoted to a source that does not make the claim.
Open questions
- The pH objection to Jordan et al. 2019 is this project's own inference and should be checked. Their vesicles need pH 11 to 12, and RNA does not survive that. If the salt tolerance and the RNA-hostile pH are the same fact, that is a significant limit on the result, and no paper states it.
- Whether anyone has run non-enzymatic RNA copying inside a coacervate-supported membrane at 50 mM magnesium. That single experiment would close the biggest gap in this article.
- The prebiotic supply of ten- to fifteen-carbon fatty acids. This is the least-discussed problem in the area and the arithmetic here is crude.
- Whether the amino-acid rescue at 10 mM magnesium is compatible with RNA assembly inside the resulting multilamellar vesicles. Nobody appears to have tried.