Contribution
The paper recasts the thermodynamics of life around gradients and exergy rather than entropy. It proposes one principle meant to hold far from equilibrium, and applies it in a single sweep from convection cells to ecosystem succession and evolution. It supplies new calculations of entropy production and exergy destruction from Silveston’s Bénard data, and proposes surface temperature from airborne thermal scanning as a measurable index of ecosystem development or “integrity”.
Key insight
Do not ask whether order is consistent with the second law; ask what an imposed gradient does to a system. The authors’ answer is that it calls up whatever structures destroy it fastest, and on that reading organisms and ecosystems are elaborate, remembered versions of a vortex in a draining bottle.
Assumptions
- The unified principle (Hatsopoulos & Keenan; Kestin): an isolated system, after constraints are removed, reaches a unique equilibrium independent of the order of removal. Kestin’s proof gives Lyapunov stability of equilibrium.
- The “restated second law” is offered as a corollary: “Implicit in this conclusion is that a system will resist being removed from the equilibrium state.” That is a local stability property of equilibrium. The step to far-from-equilibrium behaviour (“utilize all avenues”) is not derived. Le Chatelier’s principle is cited as an instance, and it too is a near-equilibrium result.
- Gradients and exergy, not entropy, are taken as the variables, so as to avoid defining entropy away from equilibrium. Yet the Bénard analysis computes an entropy production rate.
- Ecosystem development is read as the second law at work. The list of seven expected successional trends is derived from “the more energy that flows … the greater the potential for degradation”, not from a model.
- Surface temperature measures exergy degradation: with incoming K* fixed, a cooler surface emits less long-wave L* = εσT⁴, so more net radiation Rn is degraded.
Key results
- Bénard cells (Figs. 1–2, points 1–9). From Silveston’s data (silicone oil, 6.98 mm deep): heat transfer rises linearly with ΔT, entropy production and exergy destruction nonlinearly, and both jump at the onset of convection, given as Ra ≈ 1760. Nu = Q/Q_c = P/P_c = Ø/Ø_c for pure heat transfer. They reject maximum entropy production for “entropy production change being positive semi-definite as you increase the gradient”.
- Tornado in a bottle. Draining takes about 6 minutes without a vortex and about 11 seconds with one.
- Crystal River marshes (Table 1). Stressed against control: biomass −34.7%, imports −18.1%, total system throughput −20.7%, food-web cycles 142 → 69 (−51%), trophic levels 5 and 5. I recomputed the percentage changes; they match.
- Oregon forest thermal scans (Table 2, Luvall & Holbo 1989). Surface temperature falls from 50.7 °C (quarry) and 51.8 °C (clear-cut) to 24.7 °C (400-year-old Douglas fir). Rn/K* rises from 62% to 90%.
- Regional energy budgets (Table 3, Sato et al.). The evapotranspiration share is 70% for the Amazon and 2% for the Sahara. A SiB model run replacing rainforest by grassland lowers it from 63% to 48%.
- Origin of life. Life is “another route for the dissipation of induced energy gradients”; the gene is a “memory” that lets the process continue without restarting by chance. Eigen’s hypercycles are cited as the thermodynamic route.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Systems moved from equilibrium use all available means to counter the applied gradients, increasingly so as the gradient grows | assertion | offered as a corollary of Kestin’s equilibrium-stability proof; the step beyond equilibrium is not derived |
| C2 | Bénard convection raises heat transfer, entropy production and exergy destruction above conduction | strong, for these data | recalculated from Silveston’s experiment (Fig. 2) |
| C3 | The governing principle is not maximum entropy production | moderate | an inference from the Bénard data (point 9); no other system examined |
| C4 | Life arises and persists as a further means of dissipating the solar gradient | assertion | analogy from C2 and the vortex; no model |
| C5 | The second law is necessary but not sufficient for life | stated as a limit on C4 | the authors’ own caveat |
| C6 | Ecosystems develop to degrade more incoming exergy (seven predicted trends) | weak | one stressed/control marsh pair, no replicate or statistics; trends said to “subsume” Lotka, H. T. Odum, Lieth |
| C7 | More mature ecosystems are cooler and degrade more net radiation | moderate as a correlation | thermal-scanner transects and one table (Table 2); confounded with evapotranspiration, which the authors say does most of the degrading |
| C8 | Ecosystem surface temperature can index ecosystem health or integrity | proposal | ”much more research is needed” |
Concepts
- Gradient. A difference (temperature, pressure, chemical potential) holding a system away from equilibrium. The paper’s measure of distance from equilibrium.
- Exergy (availability). The available-work content of energy. Its destruction is “degradation”.
- Dissipation vs degradation. Dissipation moves energy through a system; degradation destroys exergy, the ability to set up gradients. They coincide only for pure heat transfer. The authors say Prigogine “at times mistakenly uses these terms interchangeably”.
- Restated second law. C1.
Connections
- Prigogine and the Brussels school; Paltridge; Swenson; Lotka, H. T. Odum, Jørgensen. The paper’s own lineage and foils on dissipative structures and on maximum-power and entropy-production principles. These are being filed in parallel in this batch, and nothing here relates to them by code.
- Schrödinger, LIT-530 (Deferred). The stated starting point. The paper’s reading of him (two processes, “order from order” and “order from disorder”) is unchecked here.
- Boltzmann 1886 is quoted on the “struggle for entropy” made available by the sun–earth temperature difference.
- Eigen and Ishida on hypercycles and quasi-species, quoted for the origin of life.
- LIT-036 (primitive metabolic cycles) for the autocatalytic cycles this paper invokes without modelling.
Bearing on the record
- LIT-192 / NOTE-094, the demarcation question. The paper deliberately makes the line between living and non-living dissipative structures a matter of degree (“the most sophisticated (until now) end in the continuum”). Its one qualitative marker is heredity (“memory”), and its own C5 says thermodynamics cannot supply the line. It therefore supplies no principled thermodynamic demarcation, and does not claim to.
- LIT-211 / NOTE-109. Paradigm of the “Dissipative Self-Organizing Systems” definitions.
- THEORY documents. None in the record concerns this. The record’s thermodynamic THEORYs (THEORY-026, THEORY-030) are about information and work and do not overlap.
- No ML instruction. Nothing here belongs in the anthology.
- No new THEORY is indicated. C1 and C4 are programmatic and too weakly supported to source one.
Note of 2026-10-16: THEORY-532 now cites LIT-533 as a source.
Limitations
- C1 is a proposal. Its relation to the classical second law is analogy, and the paper gives no case in which it predicts something the classical law does not.
- The physics examples are chosen because organized structure increases dissipation there. No counter-case is examined, such as a structure that forms and reduces throughput.
- The ecological evidence is correlational. Table 1 is one pair of marshes; Tables 2–3 compare sites that differ in water, canopy and climate as well as maturity.
- The paper sets aside genetics on purpose (“We have intentionally not discussed the application of these principles to … genetic processes”), so the demarcation it would need is out of scope.
Open questions
- Is there any system in which the “restated second law” makes a prediction that differs from the classical second law plus the dynamics, and is it borne out?
- Does ecosystem surface temperature track development once water availability and canopy structure are controlled?
Corrections
- none to a seeded skim (there was no seed)
- Table 2 is internally inconsistent for the 400-year forest. The table defines Rn = K* − L*. For the other four sites the figures agree (e.g. 718 − 273 = 445 for the quarry). For the 400-year stand it lists K* = 1005, L* = 95 and Rn = 830, but 1005 − 95 = 910. The 90% the text quotes is 910/1005 (90.5%); 830/1005 is 82.6%. So the Rn entry, not the 90%, is probably the misprint.
- Temperatures in the text. “The coldest site, 299°K, some 26° colder than the clear cut”: Table 2 gives 24.7 °C (297.9 K) against 51.8 °C, a difference of 27.1 °C. The text’s figures are rounded inconsistently with the table.
- The solar figure. “the 1580 watts/meter² of incoming solar energy” is higher than the usual top-of-atmosphere solar constant of about 1,360–1,370 W/m². The paper gives no source for its number.
- The onset of convection. The paper gives the critical Rayleigh number as 1760 and cites Chandrasekhar (1961). Chandrasekhar’s value for two rigid boundaries is about 1708. The paper may mean the experimental onset in Silveston’s data, but it does not say so.