. . . . "PBDB data service 1.2: ICS stage list (intervals/list.json?scale=1, type 'age'; 95 stages with duration >= 1 Myr) and, for each stage and each of six PBDB continents (cc=NOA, SOA, EUR, AFR, ASI, AUS), the marine Metazoa genus list from occs/taxa.json (rank=genus, envtype=marine, timerule=major): 570 API calls, cached and hashed individually (derived/files.json). (A) For every continent-stage with >= 100 genera (357 of them), the Jaccard similarity of its genus list with the pooled genus list of the other five continents was computed for all 95 stages and the best-matching stage recorded (derived/stage_matching.csv). (B) For every continent, proportional genus loss between consecutive stages (1 - |G_i & G_i+1| / |G_i|, both stages >= 100 genera) was computed; Spearman correlation of these series between every continent pair with >= 10 shared boundaries (13 pairs), and a permutation test of the mean pairwise rho (stage order permuted independently within each continent, 2000 permutations, seed fixed in analysis.py). Verdict follows the pre-registered rule: supported requires (A) >= 85% same-or-adjacent and (B) mean rho > 0 with p < 0.01; refuted if (A) < 60% or (B) mean rho <= 0; otherwise mixed." . . "Measured: 0.7843 fraction of continent-stages (>= 100 genera) whose fauna best matches the same or an adjacent stage on the other continents (n = 357, 95% Wilson CI [0.739, 0.824]). Interpretation: Darwin's 'forms of life change almost simultaneously throughout the world' is operationalised by its own consequence, which he states: that formations in distant regions can be correlated by their fossils. Test (A) is that correlation exercise done with genera: a continent's fauna from one stage should pick out the same stage elsewhere in the world. Test (B) checks that the amount of change between consecutive stages co-varies across continents (synchronous turnover). 'Almost simultaneously, in a broad geological sense' is read as within one stage (typically 2-10 Myr). PBDB continents are the geographic units; marine environments only. Known limitation stated up front: PBDB stage assignments rest partly on biostratigraphic correlation, which makes some agreement across continents expected by construction; a failure of test (A) would nevertheless be possible if faunas were strongly provincial, and test (B) uses magnitudes of change that are not used in dating." . "Claim test of darwin1859-ch10-12: mixed, confidence 0.6" . . . "Planned datasets: pbdb-occs-taxa-by-stage-and-continent" . "claim-test/0.1.0" . "biological" . "Pre-registered. The claim is REFUTED if the same-or-adjacent matching rate in (A) is < 60%, or the mean pairwise correlation in (B) is <= 0. Other outcomes are 'mixed'. Exact-stage matching rate, matching rate within two stages, and per-continent-pair correlations are reported as supplementary numbers." . "Darwin's claim is about marine forms changing 'almost simultaneously throughout the world'; it is here reduced to genus composition of marine Metazoa on six continental units at stage resolution (typically 2-10 Myr). It does not test synchrony at finer than stage resolution." . "Matching success depends strongly on sampling: Europe 88%, Asia 86%, North America 85% (the three continents with median > 300 genera per stage) versus Africa 66%, Australia 55%, South America 54% (median 165-185 genera per stage). The failures are mostly a sampling-depth problem rather than faunal provinciality: for the 77 mismatched continent-stages the true stage's Jaccard is on average 79% of the best stage's, i.e. the correct stage is nearly tied." . "Mismatches concentrate in the Neogene-Quaternary (e.g. Calabrian faunas of three continents best match the Aquitanian) where genus-level turnover is low and many genera persist for the whole period, and in thin Permian/Oligocene records. Low turnover is the opposite of what would falsify Darwin's claim (asynchronous change), so the 'mixed' verdict understates support for synchrony of change while correctly recording that genus lists alone cannot date stages to within one stage on poorly sampled continents." . "PBDB stage assignments of collections rest partly on biostratigraphic correlation, so some cross-continent agreement in (A) is expected by construction; test (B) uses magnitudes of turnover, which are not used in dating, and is the more independent test. The PBDB is a live database; the retrieved responses are hashed and dated so the run can be audited but later runs may differ." . "Test (A) fell short of the pre-registered 85% bar: 78.4% (95% Wilson CI 73.9-82.4%) of continent-stages are best matched by the same or an adjacent stage elsewhere (53.8% exact, 87.4% within two stages, median offset 0). Test (B) passed clearly: mean pairwise Spearman rho of proportional genus loss = 0.31, permutation p = 0.0005; 12 of 13 continent pairs positive. The pre-registered rule therefore gives 'mixed'." . "Pre-registered. Using Paleobiology Database marine Metazoa genus lists per ICS stage (stages >= 1 Myr) for each of six PBDB continents (North America, South America, Europe, Africa, Asia, Australia), (A) for every continent-stage with >= 100 genera, find the stage whose pooled genus list from the other five continents has the highest Jaccard similarity with it; and (B) for every continent, compute the series of proportional genus loss between consecutive stages (fraction of genera in stage i absent from stage i+1, both stages >= 100 genera) and the Spearman correlation of this series between every pair of continents. The claim is SUPPORTED if in (A) the best-matching stage is the same or an adjacent stage for >= 85% of continent-stages, AND in (B) the mean pairwise Spearman correlation is > 0 with permutation p < 0.01 (independent permutation of stage order within each continent, 2000 permutations)." . "Pre-registered. The claim is REFUTED if the same-or-adjacent matching rate in (A) is < 60%, or the mean pairwise correlation in (B) is <= 0. Other outcomes are 'mixed'. Exact-stage matching rate, matching rate within two stages, and per-continent-pair correlations are reported as supplementary numbers." . "0.6"^^ . . "mixed" . "Marine faunas on different continents change in step: the genus composition of a stage on one continent identifies the same (or an adjacent) stage on the other continents, and the magnitude of faunal change between consecutive stages is correlated across continents." . "PBDB data service 1.2: ICS stage list (intervals/list.json?scale=1, type 'age'; 95 stages with duration >= 1 Myr) and, for each stage and each of six PBDB continents (cc=NOA, SOA, EUR, AFR, ASI, AUS), the marine Metazoa genus list from occs/taxa.json (rank=genus, envtype=marine, timerule=major): 570 API calls, cached and hashed individually (derived/files.json). (A) For every continent-stage with >= 100 genera (357 of them), the Jaccard similarity of its genus list with the pooled genus list of the other five continents was computed for all 95 stages and the best-matching stage recorded (derived/stage_matching.csv). (B) For every continent, proportional genus loss between consecutive stages (1 - |G_i & G_i+1| / |G_i|, both stages >= 100 genera) was computed; Spearman correlation of these series between every continent pair with >= 10 shared boundaries (13 pairs), and a permutation test of the mean pairwise rho (stage order permuted independently within each continent, 2000 permutations, seed fixed in analysis.py). Verdict follows the pre-registered rule: supported requires (A) >= 85% same-or-adjacent and (B) mean rho > 0 with p < 0.01; refuted if (A) < 60% or (B) mean rho <= 0; otherwise mixed." . "Pre-registered. Using Paleobiology Database marine Metazoa genus lists per ICS stage (stages >= 1 Myr) for each of six PBDB continents (North America, South America, Europe, Africa, Asia, Australia), (A) for every continent-stage with >= 100 genera, find the stage whose pooled genus list from the other five continents has the highest Jaccard similarity with it; and (B) for every continent, compute the series of proportional genus loss between consecutive stages (fraction of genera in stage i absent from stage i+1, both stages >= 100 genera) and the Spearman correlation of this series between every pair of continents. The claim is SUPPORTED if in (A) the best-matching stage is the same or an adjacent stage for >= 85% of continent-stages, AND in (B) the mean pairwise Spearman correlation is > 0 with permutation p < 0.01 (independent permutation of stage order within each continent, 2000 permutations)." . "false"^^ . . "genus lists per stage per continent" . "CC BY 4.0" . . "Location (not a URL): https://paleobiodb.org/data1.2/occs/taxa.json?base_name=Metazoa&interval=&rank=genus&envtype=marine&timerule=major&cc=&limit=all (one call per stage and continent; per-call SHA-256 in derived/files.json)" . "Dataset: Paleobiology Database, data service 1.2: occs/taxa (marine Metazoa genera per stage and continent) and intervals/list (stage ages)" . . . "sha256:c62debddcc4cec90dd5f85fe0cd0c026354fd02f3c5d5d9273b8bc91a9596f4a" . "sha256:e3082cbcd8db6bd4011120ad17b47d81c074e66490d6d0dbc7140cf36fd37bfa" . . "delphisci claim tester (claude-fable-5.1; cursor/claim-tests-darwin-batch-b201)" . . . "claude-fable-5.1" . . . . . . . . . "2026-11-08T16:24:31.822Z"^^ . . . . . "Claim test of darwin1859-ch10-12: mixed, confidence 0.6" . . "RSA" . "MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAyFHtPB6p0mCjDozxlOY9CmO3nBm7Te3jFbXyLmM+G1tGe47Da0jYIbhkl9Oo1ZjxmCa3LCwlEqK8gj6hEIcL4ahW30JXh9HycfKVk1bugkzMEV15tJdTzAU3sGMFnbNsn43xY2LFLy45tApkRm5byEiM7oRRrooIdwoQKnP4Wzq/bfAoJ9qvW+MGkelyOTd1jbSt+KfmT7xKmDfWzAHddUlWwMX2QJOLxDaffOteTop6ZRgPiNmPJKA/S62Gac9XtSYT3oDkZ6/nr5cxxNLpOyxTf5FPISiCe11xYO/unYjY+mwnOfA8yhJIyx1l0O9I0vpOpn+E5TvyHJV+oSTrvwIDAQAB" . "CDIcGrtViTZ/4kYit19aTMEPiLGoezSYWm01yD9rFvgD3VFOw5NPuCh8n4JPXsRveMyABxumdhKpMRAKw0qHHVST+vePw2XVbwMG9t94oAmvjmFwcXXgwbZAjf622pS3J5SWmlGakukgxhjcW7FklSoWT2IC1ZFhQP8rhNXvM3AeTZqMbx2u9Idqbj/yGWUIP6ev735SJqFcQZaS+PCPpHUW5SLnPtwzmO7vyB74kSpvWIhlGJFff/HRm8Q81m/SVek6AoY9MYS3lH4HlWALMz9cUDAiZQ5tmgPDPEnhrIF04P/6IRY41fXydZXQRvXN5a9HpixjqqWzFwhJcavqLA==" . . .