Fedderson-Daedalus · Source catalogue

The file name and PDF metadata say 'Daedalus' (the Academy's later journal), but the title page shows it is the Proceedings, vol. 25, Boston, University Press, 1890, a Google Books scan of 406 pages. The contents list 22 papers, the meeting records, memoirs of eight Fellows and an index. Only two papers (VIII and XI, both by Trowbridge) concern Feddersen-type oscillatory discharges; others bearing on electricity are I, V, XVI, XVII, XVIII and XX, and the Joule memoir.

Relevance: Identifies what the file really is. It is not Feddersen's own work and not the journal Daedalus.

Mainstream conflict: none

Using a Gordon induction balance with a telephone as detector, the authors measure the specific inductive capacity (dielectric constant) of cable insulators at telephonic frequencies, motivated by an American Bell Telephone rule that line efficiency depends on the product of resistance and capacity. They find capacity varies with the rate of charging, with no general rule, and that ordinary galvanometric cable capacities predict telephone performance poorly. Digest values: petroleum 1.6, solid paraffine 2.0, indiarubber 3.7, gutta-percha 4.2, glass 4.6, water 6.8.

Key claims

• The measured capacity of an insulator depends on the rate at which it is charged, so slow (galvanometer) capacities do not predict behaviour at telephone frequencies.
• Telephonic specific inductive capacity of glass is 4.6 and of water 6.8 (digest values).

Relevance: Contemporary 1889 evidence in the same volume that capacity depends on charging rate (dielectric absorption in solids). It is the context for Trowbridge and Sabine's premise that slow-method capacities cannot be used for fast discharges.

Mainstream conflict: none; rate-dependent capacity of solid dielectrics is established (dielectric absorption).

Descriptions of fungi parasitic on insects, lichen reproduction (no sexual organs found in several genera), an alga, new plant species from Mexico and the western US. None connects to oscillatory discharges.

Relevance: None. Listed so every part of the file is accounted for.

Mainstream conflict: none

Over two years Cross measured, stroboscopically (tuning fork at 128 per second, organ pipe at 256) with a microscope, how far the electrodes of carbon and platinum microphone transmitters move. Sound is clear at imperceptible motion, scratchy at 100-200 millionths of an inch, and sparking and separation begin at 600-1000 millionths of an inch. Platinum hammers reproduce better at low amplitude; the critical excursion rises with electrode mass and with normal pressure up to a plateau; in a Blake transmitter relative excursions of 25 and 50 millionths of an inch gave good quality and strong overtones.

Key claims

• Microphone quality degrades when electrode excursion exceeds about 100-200 millionths of an inch and the contact sparks at 600-1000 millionths.

Relevance: Peripheral: electrode spark behaviour in a contact, loosely related to electrode-dependence of sparking.

Mainstream conflict: none

An improved screw compressor reaching above 2,000 atmospheres (past Andrews, and Hannay and Hogarth's 900 atmospheres), with electrically insulated piezometer tube for measurements on test samples.

Key claims

• A screw compressor with an insulated piezometer can reach and measure pressures above 2,000 atmospheres.

Relevance: Peripheral.

Mainstream conflict: none

Background: Hertz's work, Feddersen's rotating-mirror photographs of the spark (Pogg. Ann. 103-116) and Lorenz's analysis; the Thomson period t = (2 pi / v) sqrt(LC), with L inductance (cm, electromagnetic units), C capacity (cm, electrostatic units) and v about 3e10 cm/s the ratio of the units. Premise: capacity of glass measured slowly is no guide at 3 millionths of a second, so they use air condensers. Apparatus: 19 coaxial zinc cylinders (large: 15.1 to 60.4 cm diameter, 92 cm high; small: 7.60 to 25.95 cm, 30.47 cm high (the OCR and digest print 80.47; the page image at 800 dpi reads 30.47), geometric capacity 5317.9 with 'capacity of wire, 200' listed beside it), a 50 ft wire circuit with L = 39,700 cm, R = 0.742 and 1.54 ohm, a 4x5 in rotating mirror at about 54 rev/s (image speed about 1.1 mile/s), chronograph timing to 1/500 s. They tabulate oscillation periods for the small air condenser (six negatives, about 2.85 to 3.63 microseconds) and a glass Leyden jar at spark lengths of 0.4 and 1.3 cm. The large-condenser comparison has no data at all in the paper. They conclude: (1) the air condenser shows a periodicity analogous to magnetic hysteresis ('dielectric viscosity of the air'); (2) the Hertz equations do not fully describe the oscillations; (3) it is most marked when condenser capacity is suitably related to the time of the surges; (4) waves through glass in a strong magnetic field are apparently unaffected, so Maxwell's displacement current does not alter the surge time (tentative, no numbers given). Deep-read findings: the theory value 3.04 microseconds uses C = 5317.9 only; adding the listed wire capacity (200) gives 3.10, matching the right-terminal mean of 3.11 (the left mean is 3.17, +2%). Reading repeatability is about 0.063 microseconds, but spark-to-spark scatter is 2-3 times larger (0.13-0.19). A one-way ANOVA across oscillation numbers shows no reproducible pattern (right p about 0.53, left p about 0.10); the visible wave in the mean curves comes from the first ordinate. Lag-1 correlation of successive deviations is about -0.3 to -0.4, which is the alternation the authors say is 'conspicuously not the case' for setting error. The spark-length test used 1.3 cm against 0.4 cm (not 1.8 as an earlier QE note read), on a glass jar, and the two means differ by +2.2% (t about 2.3, p about 0.04), so 'no variation' is not what their own table shows; a steady resistance would need about 32 ohm for a 2% period shift (surge impedance sqrt(L/C) about 82 ohm), so spark nonlinearity rather than constant resistance is the candidate. They cite Boltzmann and Klemencic on gases only in prose with no numbers; the known departures from sqrt(K) = refractive index are for polar molecules (SO2, ethyl bromide), not for air. In the magnetic experiment (jar in a 728-turn coil at about 30 A, estimated field about 0.037 T) 'the total duration ... did not appear to be notably affected'; some runs 'seemed to show a decrease in the total number of oscillations'. Internal slips in the paper: right interval 6 of negative 1 printed 2.91 and 3.00; two printed column means do not match their columns; coil radii inverted. Second-round corrections (deep/Daedalus-Fedderson-109-c85345.md, section 10): (a) the small condenser is 30.47 cm high, not 80.47; the earlier side-check of about 11,900 cm for 18 gaps in parallel came from that misread digit and is withdrawn. With 30.47 cm, their own formula C = l / (2 ln(b/a)) summed over 18 gaps in parallel gives about 4,511 cm bare, 4,544 cm with Maxwell's end strip, which is about 15 percent below the quoted 5317.9 (the quoted value is 17 percent above the geometry). No change in the number of active gaps helps (17 gaps give 4,131). A zinc thickness of about 0.075 cm, a value the paper never gives, would close the gap; this is the deep reader's inference only. (b) With the pure geometry plus the 200 cm wire the theory period would be about 2.87 microseconds rather than 3.04-3.10, so the measured-versus-theory gap could be 8-10 percent, not 0-2 percent; the capacitance cannot be reproduced from the printed geometry to better than about 15 percent, which is a further reason the 1890 data cannot establish a few-percent departure from the LC formula. (c) The concave mirror radius is 313 cm (OCR printed 813); with the spark at 230 cm this images it at 489.7 cm, consistent with the printed layout. (d) The paper gives symbol beta for the zinc thickness but no value, and no count of active gaps.

Key claims

• Electrical oscillations in air between condenser plates show a periodicity, the analogue of magnetic hysteresis, from energy spent overcoming a 'dielectric viscosity' of the air.
• Measured periods in an air condenser deviate from t = 2 pi sqrt(LC); the authors attribute this to energy transformations in the dielectric (their numbers agree with theory to 0.3% on the right terminal once the stated wire capacity of 200 cm is included). The capacity itself is only reproducible from the printed geometry (30.47 cm height) to about 15 percent, so the few-percent comparison is not secure.
• A change in spark length from 0.4 to 1.3 cm in a glass jar leaves the period unchanged (their table shows +2.2%).
• The periodicity is less marked with a larger, slower air condenser (stated only; no data in the paper).
• Waves through glass in an intense magnetic field, with the electric strain perpendicular to the magnetic, are not noticeably affected, so Maxwell's displacement currents do not alter the surge time (tentative, qualitative).

Relevance: Central paper for claims 2 and 3 of this project. It is the only primary experimental evidence for the 'dielectric viscosity of air' idea and gives the 1890 experimental basis for Feddersen-type period measurements.

Mainstream conflict: The 'dielectric viscosity of air' explanation conflicts with the modern view that air is a negligible-loss, frequency-independent dielectric at these frequencies [memory, high]; the period scatter is better explained by the spark as a nonlinear, time-varying element and by stray capacity. The displacement-current null is qualitative and underdetermined.

Organic reaction syntheses with melting points and combustion analyses.

Relevance: None.

Mainstream conflict: none

Trowbridge asks whether an oscillatory spark carries metal atoms to and fro or only shakes them. Rowland concave grating of 20,000 lines to the inch, iron terminals, a 6 mm gap, one terminal hollow and placed perpendicular to the slit so the oscillation lies along the line of sight, with a comparison exposure (terminals parallel to the slit) on the same plate and a movable shutter. Two self-inductions gave double-oscillation durations of t = .0000003 s (3.0e-7 s; the OCR text layer reads it .0000008, an error) and t = .0000024 s (2.4e-6 s), calculated from t = 2 pi sqrt(LC); no L or C values are given, and a revolving mirror only showed qualitatively that the discharge was oscillatory. He states the grating 'enabled me easily to detect a movement of 1/100 of a wave length' (page image, about 85 percent confidence in the reading 1/100 rather than 1/1000; the OCR shows garbage). This is a claimed detection limit, and its unit is ambiguous: 1/100 of a wavelength as a literal line shift would correspond to delta/V = 0.01 (3,000 km/s), a poor test, so he probably means a fraction of a line width or spacing; the paper does not say. He says convection of iron atoms across 6 mm would give a broadening that 'might amount to a space equivalent to a whole tenth meter' (1 angstrom). The deep reader's check: 6 mm in half of 3e-7 s is 4e4 m/s, Doppler fraction 1.3e-4, about 0.5-0.6 angstrom near 4000-4400 angstrom (same order, smaller than his figure); for 2.4e-6 s about 0.07 angstrom. Result: 'The iron lines in the two photographs met exactly, continued in an unbroken line across the double photograph, and were of the same breadth', photographed near the great H lines. Conclusion: oscillations do not carry the atoms of metals with them; 'the atom is merely shaken up'. Trowbridge himself notes convection 'would take place too slowly to give a spectroscopic effect', and cites J. J. Thomson (Phil. Mag., Aug 1890) that charged atoms could not carry the discharge, with 80,000 V across 1 cm of air. The null is qualitative (visual match, no repeats reported), and the test addresses bulk atom convection, not the electron-carrier question. Published at Jefferson Physical Laboratory, presented 8 Oct 1890; the Rumford Fund gave him $500 'for metallic spectra' (meeting records p.335-336).

Key claims

• In an oscillatory iron spark across a 6 mm gap the metal atoms are not carried to and fro by the oscillation: the iron lines near the H lines show no displacement or change of breadth between the along-slit and across-slit exposures, at a claimed sensitivity of '1/100 of a wave length' (unit ambiguous).
• The oscillations shake the atoms so that they radiate rather than convey them (qualitative, single comparison, no repeats reported).

Relevance: Central paper for claim 3; applies Feddersen-type oscillation to atomic motion.

Mainstream conflict: none: consistent with the accepted view that line broadening in sparks is mainly Stark and pressure broadening, not oscillatory Doppler [memory, medium; not web-checked].

Precision analytic chemistry (Becker balance, copper by electrolysis, bromine as AgBr) and fractional distillation of sulphur compounds in petroleum (methyl and ethyl sulphides etc.).

Relevance: None.

Mainstream conflict: none

A 4,015 g curved scale-shaped meteoric iron found 1885 near Jamestown, ND; 90.24% Fe, 9.75% Ni, 0.05% P, trace Cu, no sulphur.

Relevance: None.

Mainstream conflict: none

Induced current in magneto telephone transmitters rises with magnet strength to a maximum then falls (saturation), the maximum at stronger fields for thicker diaphragms; receiver tests with a Norway iron core and ballistic galvanometer.

Key claims

• Induced current in a magneto telephone first rises and then falls as magnet field strength increases, because of saturation.

Relevance: Peripheral, magnetic-circuit context.

Mainstream conflict: none

A bench-level comparison of dielectrics for large condensers, chiefly whether cheap paraffined-paper condensers can match mica (Charles Nott, Jefferson Physical Laboratory, presented 12 Nov 1890; journal pp.244-258). Twenty condensers (Nos. 1-18, G, and the Elliott mica standard No. 72, nominal 1 microfarad) built from tinfoil 0.00066 cm thick and paper, hard rubber or mica sheets; four gravity cells in the insulation test; a Peirce ballistic galvanometer (needle period 583 s in the insulation tests, 45.1 s otherwise); a pendulum key for charging times from about 1 ms. It never uses the words absorption, hysteresis or viscosity, and measures only solids, nothing on air. Results: (1) insulation: 'apparent loss' after 14,400 minutes (10 days) is 97-98 percent for the rag-paper Nos. 3 and 6, while hard rubber No. 9 loses about 35 percent in 11,168 min; the table cannot separate conduction from absorbed charge. (2) Residual discharges: every condenser returned a residual of about 2-27 percent of the charge (best 2-3 percent for No. 3 and a short sample of No. 72; old bond-paper G 27-68 percent), in instalments that decline with repeated 2-minute discharges and recover with waiting (No. 72: 0.1 at 2-min intervals, then 1.2 after 6 min, 1.5 after 22 min, 3.0 after 28 min): the signature of slow dielectric absorption, which Nott records but does not name. 'I could not find a time of charging that would make the residue disappear' (p.293). His verdict: paraffine-paper condensers 'gave larger residual discharges than either the mica or hard rubber condensers. That was their worst fault' (p.296). (3) Capacity against charging time (about 1 ms to 30 s): leaky condensers rise 16-85 percent (No. 1 +27, No. 2 +16, No. 10 naked mica +19, No. 18 +85), good ones 1.6-5.5 percent (No. 9 +2.9, No. 11 +1.7, No. 12 +1.6, Nos. 14 and 15 +3.5, No. 72 +5.5); 'The time curve of a condenser is one of the best tests of its goodness' (p.293). The passage the earlier reviews quoted is exact: 'When as short a time as 0.002 sec. was tried in charging Nos. 14 and 15, the discharge exceeded the charge, but the results were too irregular to warrant any definite conclusions' (p.290); it compares a charging throw through the galvanometer with a later discharge throw, gives no numbers or explanation, and the only tabulated 'apparent gain' (hard rubber No. 9 after 'soaking', 28.0 to 28.1) is 0.5 percent, within scatter. Extras: commercial tissue paraffine condensers of 2.5 microfarad lost their charge in under ten minutes; a 0.808 microfarad one measured 740 ohm; a linseed-oil paper condenser acted as a chemical cell (about 1/14 volt); cost per microfarad: hard rubber $32.25, mica $11.60, paraffine paper $2.75. Scan caution: the best paper condenser is No. 3 (the OCR prints No. 8) and the hard-rubber capacity is 0.389 microfarad (OCR .889). Limits: times from about 1 ms (kilohertz, not megahertz), voltage about 4 V, no temperature or humidity, and no air; so it supports slow polarisation in solids (relevant to the glass jar in Trowbridge and Sabine), not a 'dielectric viscosity of air'.

Key claims

• Capacity measured by ballistic throw depends on the duration of charging: from about 1 ms to 30 s leaky paper condensers change by 16-85 percent and good ones (hard rubber, mica, harder-paraffine paper, No. 72) by 1.6-5.5 percent.
• Every condenser tested, including mica and hard rubber, returned a residual discharge of about 2-27 percent of the charge in successive instalments, recovering with waiting time, and no charging time made the residue vanish; paraffine-paper condensers had the largest residuals.
• At 0.002 s charging of Nos. 14 and 15 the discharge exceeded the charge, but the results were too irregular for any conclusion (no numbers given; absorption is the reviewers' reading, not Nott's).
• A paraffine-paper condenser can be made that compares favourably with a good mica one (holds for Nos. 14 and 15; the table for No. 3 shows about +27 percent variation, so that case needs a check).

Relevance: Contemporary, quantitative evidence of rate-dependent capacity and dielectric absorption in solid dielectrics, relevant to the glass jars in Trowbridge and Sabine and a plausible reading of an apparent discharge-exceeds-charge effect. It says nothing about air or megahertz oscillations.

Mainstream conflict: none

EMF of Zn/Pt, Zn/Cu and Daniell cells measured at 10-1800 atmospheres and 15-80 C; the digest gives a Daniell-cell pressure coefficient at 15 C but its exponent is garbled, so no value is quoted. Pressure and temperature coefficients have the same sign.

Key claims

• The EMF of a voltaic cell changes with pressure, with a coefficient of the same sign as its temperature coefficient.

Relevance: Peripheral.

Mainstream conflict: none

A. E. Dolbear, presented 14 Jan 1891 (journal pp.271-277). Purely magnetic and mechanical: no wireless, electrostatic, spark or speed content appears; the only distances are a galvanometer 'at such a distance that magnetic fields external to its circuit could not act upon it', a telephone in another room, and a second ring heard at 'a yard or more'. Fifteen experiments: (I-II) an 8 inch, 1 inch thick solid iron ring with a coil on a far galvanometer; a flexible iron-wire ring forced from one ellipse to another gives a deflection whose sign depends on the phase and reverses when the ring is turned over, with substantially the same result with no battery current (residual magnetisation suffices); (III-X) struck rings, one with 600-700 turns of No. 32 wire and another with about 50 turns, heard in a telephone, louder when magnetised (a 200 turn coil), audible from a second unwired ring brought near and out to a yard, louder when loaded to the same pitch (his 'sympathetic' against 'forced' vibration), and a horseshoe magnet with about fifty ohms of No. 32 wire; (XI) a coil on a 4 ft gas pipe struck two or three times a second changes loudness when rotated from the magnetic meridian, a plain Earth's-field effect; (XII) the one quantitative run: a 1 ft ring of soft iron wire, 4.6 ohm, probably 1,000 turns (his estimate), magnetised by three secondary cells (6 V, about 1,300 ampere-turns; 6/4.6 = 1.30 A, consistent), 0.67 ohm galvanometer, deflections of 30-40 degrees for large changes; (XIII) four currents per cycle, two direct and two reverse; (XIV-XV) blows or jerking demagnetise, a heavy 8 by 1.5 inch ring giving a wide swing at the first blow then less; welded steel rope rings (Elihu Thomson), 5-6 inches of movement, 400-500 turns of No. 22 wire. No control with an unmagnetised ring in the galvanometer runs, no deflection sizes except 30-40 degrees, no timings, frequencies or calibration. Four conclusions are said to be 'proved' (form change alters field stress; vibrations produce periodic field change; vibrating magnets act on other magnets sympathetically or forced; induced currents in surrounding circuits); the third goes beyond the data, as magnetic coupling is not separated from mechanical or air-borne coupling. The speculative part: atoms are elastic magnets, probably rings ('a ring magnet is the most perfect form possible'), whose periodic form change sends waves in the ether that are heat and light, consistent with Maxwell; a closed ring should stress the external ether, and the waves 'would not necessarily give rise to electro-magnetic effects directly'; this suggests the vortex-ring atom 'of which I have so often spoken to the Academy'. He says the experiments might have 'a different interpretation', and proposes no test of the ether-wave idea.

Key claims

• Changing the form of a magnetised iron ring induces a current in a coil around it (deflection of 30-40 degrees on a 0.67 ohm galvanometer at about 1,300 ampere-turns), reversing with the phase of the change; residual magnetisation alone suffices.
• Struck magnetised rings are heard in a telephone circuit at about a yard from a second ring, and rings of the same pitch respond more loudly (sympathetic vibration), which Dolbear reads as magnets acting on one another through the field (mechanical and air-borne coupling not excluded).
• Atoms are probably small, highly elastic magnetic rings whose vibration sends periodic waves through the ether, which are heat and light (speculation, no test proposed).
• A closed ring magnet would stress the external ether and emit waves without direct electromagnetic effect (stated as supposition; not measured).

Relevance: Most alternative-science content in the folder (ether, vortex-ring atoms, non-electromagnetic magnetic waves). Not part of the Feddersen oscillation chain but the best 'also in the folder' item for the reader.

Mainstream conflict: The induced currents are expected from magnetostriction (Villari effect) and flux change in a moving magnetic circuit [memory, high; this explanation is the reviewers', not Dolbear's]. The ring-atom, ether-wave and non-electromagnetic wave speculation conflicts with later physics (no ether; magnetic waves are electromagnetic waves).

Rumford Committee grants including $500 to Trowbridge for metallic spectra; Dr. Harold Whiting's air thermometer for heat from alternating current through the human body; papers presented by title; deaths of Joule and others; election of W. R. Livermore. Per QE (OCR p.344) Dolbear spoke on vortex rings at a meeting; not confirmed in the digest.

Key claims

• The Academy's Rumford Fund supported Trowbridge's spectral work (so funding was not withheld).

Relevance: Records that Trowbridge's work was funded, and flags Dolbear's vortex-ring talks.

Mainstream conflict: none

Memoirs of an MIT professor, an antiquary, a naval mathematician, an industrialist-philosopher, a historian (Johnston), the meteorologist Elias Loomis (terrestrial magnetism, aurora, weather maps), the astronomer Maria Mitchell and the clergyman Woolsey. Loomis's aurora and magnetism work is the nearest to this project's field.

Relevance: Marginal.

Mainstream conflict: none

Memorial notice, written about 14 months after Joule's death on 11 Oct 1889. Joule: born Salford 25 Dec 1818, brewery family, taught by Dalton until 1837, began investigations in 1838; about 100 papers alone and 20 with Thomson, Playfair and Scoresby. Work: quantitative electromagnetic-engine measurements (1838 on); the heating law, 'proportional to the resistance of the conductor multiplied by the square of the electric intensity' (printed RC^2, 1840), independent of shape and metal; magnetostriction (1/30,000 inch in 2 ft, five Grove cells, 1841); a zinc battery compared with a coal-fired steam engine (1839-40), a verdict the author says still holds and which bears on over-unity hopes; 'The Mechanical Value of Heat' (1843) with the first equivalent, 838 foot-pounds per pound per degree Fahrenheit (the author says 838 is too large; Joule in 1881 attributed the excess to heat absorbed by the magnet), then 781.8 (1847), 772 (1849) and 772.55 (1878). Reception: 'Some of the members of the Royal Society expressed the opinion that it was nonsense from beginning to end ... and as a body they refused to publish it in their Transactions'; Thomson in 1847, not wholly persuaded, was the only friendly voice; 'the paper was not published by the Society until 1850, seven years after its presentation'; Roscoe: 'the men who so bitterly opposed it ought to have known better'. A footnote says a British Association committee in 1849 printed a paper except the proposition 'Friction consists in the conversion of mechanical power into heat.' Known in Europe by 1852, honoured abroad before at home; 1864 British Association electrical-units committee (his value 'nearer right' than the new ohm-based one); 1878 pension of 200 pounds a year. Cautions from the deep read: the memorial dates the Cork paper to 21 Aug 1848 but also says 1850 was seven years after presentation (so 1843), an internal inconsistency in the memorial itself, not a digest error as earlier thought; it blends refusal with delay, names no objectors except Roscoe, cites no source for the quotations, is written 40-47 years after the events in a hagiographic tone, and so is a secondary report; it conflates the equivalent J with the later unit joule; and it mentions no priority question (Mayer, Colding, Helmholtz, Carnot are not named). Its list of resisted innovators (Galileo, Newton, Laplace, Young, Faraday) is the author's rhetoric. Academy roll at 15 Feb 1891: 186 Resident Fellows, 98 Associate, 74 Foreign Honorary.

Key claims

• Joule's mechanical equivalent of heat fell from 838 (1843) through 781.8 (1847) and 772 (1849) to 772.55 foot-pounds per Fahrenheit degree per pound of water (1878); the memorial explains only the 838 excess (heat absorbed by the magnet, 1881).
• Per this 1890 memorial (a secondary report, not checked against the Royal Society record), members of the Royal Society called Joule's paper 'nonsense from beginning to end' and as a body refused to publish it in their Transactions; it appeared only in 1850, seven years after its presentation, with Thomson the only friendly voice.
• In 1849 a British Association committee printed a Joule paper except the proposition 'Friction consists in the conversion of mechanical power into heat'.
• Joule found heat proportional to R C^2, independent of the shape and metal of the conductor (1840).

Relevance: A documented case of establishment resistance to a correct result, a mainstream counterweight for the alt-science reader.

Mainstream conflict: none (established physics).

Membership list and an alphabetical index of species, compounds and subjects.

Relevance: None.

Mainstream conflict: none

Question: does electrode material change how readily an oscillatory spark starts? Earlier null results: Righi 1876, De La Rue and Muller 1878, Peace 1892, Carr 1903; for polarity effects Guthe 1901, Hobbs 1905, Almy 1910. Apparatus: a gap of mechanically alike, chemically different electrodes (Cu, Fe, Zn, Bi) in series with L and C, connected across a leaky transformer secondary on 110 V 60 Hz mains; the spark is photographed on a moving glass plate (spiral of spark trains) or on a film on a 40 cm wheel at about 2,000 rpm (film speed about 42,000 mm/s, 20 turns per film). Film circuit: coil of about 100 turns of coppered iron wire, L = 25 mH (low-frequency), R = 6.5 ohm; condenser of 48 glass panes with tin foil, 'absorption was very bad', C = 1/8 to 1/2 microfarad depending on charge and discharge time. Gap 'varied from 0.1 mm to 3 or 4 mm', no gap stated for any table; applied voltage, electrode shape and surface preparation not given. Counts per half-cycle, favoured polarity vs other (trains per half-cycle): Cu-Fe 205 trains in 30 half-cycles vs 101 in 32 (printed 6.2 vs 3.16; 205/30 is 6.83, a printing slip); Fe-Bi 5.21 vs 2.86; Cu-Bi 2.16 vs 1.06; Fe-Zn 3.1 vs 1.41; Cu-Zn 4.77 vs 1.27 (the 1.27 is per active half-cycle: 14 trains over 32 half-cycles, 21 blank, is 0.44; the asymmetry is about 11 to 1). Control Cu-Cu: 63 trains in 31 half-cycles vs 74 in 33 (2.03 vs 2.24), but with about 4 sparks per train against 2-3 in Cu-Fe, and 71.5% vs 86.5% even trains, so the control was not run under matched conditions and odd/even counts are biased even for like metals (Rich notes this himself). Each pair is a single 'typical' film of about 30 half-cycles per polarity; no statistics. No Bi-Zn data exist anywhere in the paper, so the ranking Fe, Bi, Zn, Cu does not directly place Bi against Zn (it leans on Guthe's coherer series). Polarity wording is mixed: the table headings and the closing sentence ('electrons are emitted from iron') say the favoured electrode is the NEGATIVE one, but the prose repeatedly says 'iron anode', 'bismuth anode', 'zinc was anode' for that same column. Rich's own spark spacing of 4.1 mm at 42,000 mm/s is 97.6 microseconds; with one spark per half-cycle the period is about 195 microseconds and the effective capacity about 0.039 microfarad (or 0.0097 if one spark per cycle), 3 to 13 times below the quoted low-frequency 0.125-0.5 microfarad, consistent with dielectric absorption in the glass condenser and eddy loss in the iron-wire coil. Rich cites Schuster-Hemsalech and Milner and Royds on metal vapours leaving different electrodes at unequal speeds. Printed slips: Cu-Cu even-side sparks 291 vs 292 from its own rows. Remaining pages are American Physical Society officer lists and covers.

Key claims

• An oscillatory spark between electrodes of different metals starts more readily on one polarity than the other, so the gap partly rectifies the alternating supply.
• The favoured polarity is the one with the higher-ranked metal negative, in the order Fe, Bi, Zn, Cu (Bi against Zn is not measured), which Rich reads as ease of electron emission.
• With identical electrodes the spark trains are roughly symmetrical (Cu-Cu: 2.03 against 2.24 trains per half-cycle, with a stated bias in odd/even counts).
• Electrode material is not negligible in spark initiation in air at atmospheric pressure, contrary to earlier null results.
• The effective capacity of Rich's glass condenser at the oscillation frequency is about 0.04 microfarad, several times below its low-frequency value (QE calculation from Rich's 4.1 mm spacing, confirmed by the deep reader).

Relevance: The only experimental paper in the folder; principal evidence for the project's testable claims (4 and 5).

Mainstream conflict: Partly compatible: dependence of breakdown and arcs on electrode material and polarity (work function, cathode fall) is accepted. Departs from the older null results at ordinary gaps. Single films, no gap values, no controls for oxidation, shape or supply asymmetry, and a control run at different damping weaken it [reviewer judgement].