◐Partially reproducedMethods, Paleointensity Selection criteria, paragraph 2; Table 2rests on an assumption
For the nine failed glass specimens with a line fit (ss19, ss43, ss22, ss41, ss6, ss9, ss10, ss30, ss34), the number of points, temperature range, MAD, R2 and pTRM check differences are as in Table 2, which lists the criteria each failed.
Reasons for de-selection are provided in Table 2
Methods, Paleointensity Selection criteria, paragraph 2; Table 2
With the mean vector of each step's repeated measurements, n matches Table 2 for all nine specimens, and 23 of 29 values match under one of the named readings (79%, the criterion asks 80% for reproduced and 50 to 80% for partially reproduced), MAD 8 of 9, R2 6 of 9, pTRM checks 9 of 11. Measured values are MAD under reading (a), a free line, and R2, in the claims file's order; the last value is the fraction of the 29 values that match. MAD reading (a) alone gives all 8 matching MADs (median difference 0.2 deg, six within 0.3 deg); readings (b) and (c) give 2 and 0. Outside tolerance under every reading are ss10's MAD (33.3 against 42.9 deg) and R2 (0.23 against 0.15), ss9's R2 (0.89 against 0.74), ss34's R2 (0.03 against 0.06, tolerance 0.006), and the pTRM checks of ss19 at 250 C (23.7 to 105.9% against 30%) and ss30 at 450 C (8.2 to 43.8% against 10%). The matching pTRM checks need different readings for different specimens; no single reading matches more than 5 of 11. With the first measurement of each step, 11 of 29 values match (38%), which would not reproduce the claim. Against the stated thresholds, reported and not judged, ss30 and ss34 have R2 below 0.9 and n of 4 or more, as their single failure "R2" says; ss6 has R2 0.89 here, so its listed R2 failure follows here. Counting ss6's R2 as a miss, since its 250.1 step averages two groups of directions, gives 22 of 29 (76%). With one pTRM reading held fixed the score is 17, 16, 19 and 14 of 29 (readings a_i, a_ii, b_i, b_ii; 48 to 66%), partial or lower under every fixed reading. None of the nine is a synthetic sheet. Within a step, ss10's field-off steps and its 100, 175 and 250 C field-on steps scatter by 20 to 113 deg (resultant over n 0.79 to 0.98), and ss9's 360 and 550 C steps by 57 and 30 deg, so how these steps are combined moves the MAD and R2 that miss. The deposit's Supplementary Figure 1, read from its vector drawing, plots TRM gained from the mean vectors and NRM as the mean of the rows' moments for 12 of its 14 non-synthetic specimens (median difference 0.1% or less; ss27 and ss34 fit no reading), including ss6's averaged 250 C step, and 11 of its 12 pTRM checks at the check's TRM referred to the field-off step before it (reference i). With NRM read that way, R2 matches for 7 of 9 (ss10 0.152 against 0.15; ss9 0.916 and ss34 0.032 still miss) and 24 of 29 values match under any reading (83%); holding pTRM reference (i), as the figure plots the checks, gives 8 of 11 pTRM checks and 23 of 29 (79%). The outcome is kept, as the figure's NRM reading crosses the 80% bar only together with a pTRM reference the same figure does not use.
How the test differed from the paper’s method
- The paper does not say how the repeated measurements of a step (2 to 19 per step) enter Table 2. Assumed here that they are averaged as vectors, the usual treatment of repeated measurements and the reading under which the MADs of six specimens come within 0.3 deg of Table 2. The first measurement of each step is shown as the other reading and gives a different outcome. The deposit's Supplementary Figure 1 supports averaging; it takes NRM as the mean of the rows' moments, not the moment of the mean vector, which changes R2 by at most 0.08 here and is shown as a second reading.
- MAD is Kirschvink's principal component measure under the claims file's three readings, and the pTRM check difference is taken on the TRM-gained axis, where Figures 1 and 2 plot the checks (legends, "red triangles represent partial thermal remanent magnetization checks"). The paper gives no formula for either; it names PmagPy and "standard paleointensity definitions" (Technical Validation).
- ss6's 250.1 step holds 19 rows in two groups of directions (inclination near 71 and near -13 deg); all are averaged, as for every step. Ten other steps of six glass sheets also hold two groups of directions more than 45 deg apart (s11 360.1, 550 and 550.1, ss10 250 and 360, ss18 100, ss24 NRM and 550, ss29 175.1, ss36 450.1) and are averaged the same way.