The Sea's Missing Salt

"Our result is an upper limit on the age of the ocean.  It does not mean that the true age is anywhere near 62 Myr. ...  Our conclusion from the Na+ data is that the sea is less than 62 million years old.  This is at least 50 times younger than the age evolutionists require it to be." — Steven A. Austin and D. Russell Humphreys, Institute for Creation Research, 1990¹

"If it be objected that the Water of the Ocean, and perhaps of some of these Lakes, might at the first Beginning of Things, in some measure contain Salt, so as to disturb the Proportionality of the Encrease of Saltness in them, I will not dispute it:  But shall observe that such a Supposition would by so much contract the Age of the World, within the Date to be derived from the foregoing Argument ... though perhaps by it the World may be found much older than many have hitherto imagined." — Edmond Halley, Royal Society, 1715, in the paper that invented this argument²

"Are the deep sea deposits of salt as large as we think, or are there possible other deep sea deposits that remain undiscovered?  Is there an unknown mechanism for regulating the salinity of the ocean?  The answers to these questions must lie in innovative new approaches to understanding the history of seawater." — William Hay and five colleagues, closing the best modern reconstruction of ocean salinity, 2006³

Summary

The ocean holds about 1.47 × 10¹⁹ kilograms of dissolved sodium.  Rivers, groundwater and the sea floor deliver more of it every year.  Known processes remove only a fraction of what arrives.  Divide the amount in the sea by the net gain and you get a number of years, and the number is tens of millions, not billions.  Edmond Halley proposed the method in 1715.  John Joly carried it out in 1899 and got 89 million years.  Steven Austin and Russell Humphreys redid it in 1990 with modern flux data, added a sink inventory and a differential equation, and got a maximum of 62 million years.¹

1. Their arithmetic is correct, and this article is the place that settles it. All five published results reproduce to three significant figures from their own two tables, using a script written for this article. There is no algebra error, no dropped factor, no cherry-picked exponent. Anyone who rebuts this argument by alleging a calculation mistake is wrong.
2. The answer is controlled by two numbers, and both are weak. Delete the continental-groundwater input, which rests on a stipulated sodium concentration the authors introduce with the words "we know this average imperfectly," and the ceiling moves from 62 million years to 109 million. Published estimates of the mid-ocean-ridge sodium sink span a factor of forty, and at the top of that range the bound vanishes entirely.
3. Their second-largest input comes from a table its own authors disown. A quarter of their total sodium input is taken from Drever, Li and Maynard's 1988 Table 1.4. Three pages earlier those authors write that the table's magnitudes are "too high by a factor of at least 2," and that the sodium entry in particular is probably unwashed sea salt.
4. The measured record of seawater runs the wrong way. Fluid inclusions trapped in ancient halite are samples of real seawater, and they show sodium going up and down rather than steadily up: relatively high in the Permian, relatively low in the Cretaceous, high again today. A quantitative reconstruction of ocean salinity from evaporite volumes finds a general decline across the Phanerozoic, from the low 50s per mil in the Early Cambrian to 34.7 today.
5. The standard rebuttal is also wrong, and the article says so. Glenn Morton's 1996 counter-budget, and the TalkOrigins entry built on it, balance a maximum-past output column against a minimum-past input column. Corrected, the ratio is 0.833, still a deficit. Strip out Morton's one invented term and it falls to 0.754. There is a genuine unclosed sodium budget in the mainstream literature, and pretending otherwise is not honest.
6. The biggest defect was put to Austin in print, at a creationist conference, four years before he published. Morton raised the buried-rock-salt objection at the First International Conference on Creationism in 1986. The proceedings printed Austin's reply, which is a bibliography, and Morton's closure: "It is difficult to reply to a bibliography."
7. The argument most readers have met is not this argument. Henry Morris's Scientific Creationism runs a twelve-element "shortest residence time wins" calculation headlined by nickel at 9,000 years, and its own printed table gives lead 2,000 years against sodium 260,000,000. That version refutes itself on the page. Austin and Humphreys' version does not, and grading both by the same standard means saying so.

The verdict.  The calculation is sound and the conclusion does not follow.  What the sodium data actually show is that the ocean is not in steady state, which mainstream geochemistry has said in print since at least 1988.  The mainstream answer is a fluctuating ocean, not a constant one.  Austin and Humphreys are right that steady state is false.  They are wrong that not-steady-state means monotonically rising, and the halite inclusions say so directly.

Every source this article rests on is catalogued, with what it measured and which way it cuts, in a companion page: the literature.  Six quotations and figures that could not be verified against the source they are credited to are kept, unprinted and undeleted, in quarantine.

Where the argument comes from

Halley, 1715: the method, and three caveats its users drop

Edmond Halley read a five-page paper to the Royal Society in 1715 with a title that says exactly what it does:  "A Short Account of the Cause of the Saltness of the Ocean, and of the Several Lakes That Emit No Rivers;  with a Proposal, by Help Thereof, to Discover the Age of the World."²  Rivers carry salt to the sea.  Water leaves by evaporation and the salt stays.  So the sea gets saltier, and the rate at which it gets saltier is a clock.

Austin and Humphreys credit him in their first paragraph, and the credit is accurate as far as it goes.¹  But the paper contains three things that cut against every modern use of it, and a young-earth encyclopedia that quotes Halley as the founder of this argument while omitting them is not quoting him honestly.

Halley did not read Genesis 1 as six ordinary days, and said so on the first page.  His reason was internal to the text:

"But whereas we are there told that the Formation of Man was the last Act of the Creator, 'tis no where revealed in Scripture how long the Earth had existed before this last Creation, nor how long those five Days that preceded it may be to be accounted;  since we are elsewhere told, that in respect of the Almighty a thousand Years is as One Day, being equally no part of Eternity;  Nor can it well be Conceived how those Days should be to be understood of natural Days, since they are mentioned as Measures of Time before the Creation of the Sun, which was not till the Fourth Day."²

His method is a differential measurement, not a division.  He does not propose measuring the salt once and dividing.  He proposes measuring it, waiting centuries, and measuring it again:

"for if it be observed what Quantity of Salt is at present contained in a certain Weight of the Water of the Caspian Sea, for example, taken at a certain Place, in the dryest Weather;  and after some Centurys of Years the same Weight of Water, taken in the same place and under the same Circumstances, be found to contain a sensibly greater Quantity of Salt than at the time of the first Experiment, we may by the Rule of proportion, take an estimate of the whole time wherein the Water would acquire the Degree of Saltness we at present find in it."²

And he is explicit that this makes his own argument useless to him:  "But since this Argument can be of no use to Ourselves, it requiring very great Intervals of time to come to our Conclusion, it were to be wished that the ancient Greek and Latin Authors had delivered down to us the degree of the Saltness of the Sea."²

He raised the initial-salt objection himself, and answered it the same way modern young-earth writers do.  That is the last paragraph of the paper, quoted at the top of this article.  A sea that started salty makes the derived date a maximum, not an estimate.  Three hundred and eleven years later, that is still the entire logical content of the argument.

Halley's target was not an old earth.  It was the eternity of the world, the Aristotelian doctrine that the cosmos had no beginning, and he expected the salt clock to make the world "much older than many have hitherto imagined."²

Joly, 1899: the argument in almost its modern form

John Joly measured the sodium in nineteen rivers, estimated the ocean's bulk from fathom lines, divided, and published the answer in the Scientific Transactions of the Royal Dublin Society.⁴  He states the plan in his opening paragraph:

"Such an element is sodium.  We take for this calculation the element alone, thus avoiding the obscure question of its ionisation, which does not concern the issue.  The quantity of sodium now in the sea, and the annual rate of its supply by the rivers, lead, it will be seen, to the deduction that the age of the earth is 99 × 10⁶ years.  Certain deductions from this are, it will be shown, warranted, so that the final result of this paper will be to show that the probable age is about 89 × 10⁶ years."⁴

His numbers hold up remarkably well.  Recomputed against the values Austin and Humphreys use ninety-one years later, Joly's ocean sodium inventory is 8% high and his net river sodium flux is 19% low.⁵  Austin and Humphreys call it "extraordinary accuracy" and the compliment is deserved.  This is real quantitative geology done with fathom lines and wet chemistry.

Almost the whole skeleton of the 1990 argument is Joly's, and that should be said plainly.  Inventory over flux;  the answer read as a maximum rather than an estimate;  the explicit warning that recycled sea-spray sodium must be subtracted first.  Joly's own verdict:  "The least speculative part of our knowledge inclines us to believe that this is probably a major limit."⁴  What Austin and Humphreys add is eight further input terms beyond rivers, a tabulated inventory of sinks, and a differential equation that converts a filling time into a maximum age when the sinks are non-zero.  Their answer is smaller than Joly's mostly because of the extra inputs, not because of the equation.

They also drop Joly's most interesting caveat.  Joly did not assume a sodium-free primeval ocean.  He assumed a primeval ocean of chlorides and deducted the sodium released when free acid first attacked the primeval rocks:

"We have to consider, indeed, in this matter that the ocean was not always charged with its present dissolved salts.  The primeval ocean, most probably after the free acids were satisfied in the solution of the silicates, carried chiefly chlorides indeed, but chlorides of lime, magnesia, and other metals."⁴

That deduction is 1,972 × 10¹² tons of sodium, about 13% of his ocean total.  Austin and Humphreys set the initial sodium to exactly zero instead, and are open about why:  it makes the derived age as large as possible, which is the conservative move when you are computing an upper bound.  That is correct reasoning.  It is still worth noticing that the man who first executed the calculation thought the zero-initial assumption was wrong.

1966 to 1987: the young-earth literature rejected this clock

A straight line from Halley to Joly to Austin and Humphreys is not the line the young-earth movement actually walked, and the detour is worth reporting because it cuts both ways.

Melvin Cook rejected the ocean-salt clock in 1966, twenty-four years before Austin and Humphreys revived it.  Cook was a professor of metallurgy at the University of Utah and the movement's leading technical author of the period;  Henry Morris and Harold Slusher both cite him as an authority.  His Prehistory and Earth Models devotes a section to sodium in the oceans and opens it by conceding the point:

"While the arguments of salt balance appear adequate to show that the salt content of the oceans will not build up significantly in geological time, the problem of sodium should still be considered.  All cations in the oceans must be in complete cation exchange equilibrium with the deep sea sediments or clays.  This fact was recognized by Goldschmidt [1954]."⁶

His mechanism is ion exchange on deep-sea clays, which is Austin and Humphreys' output term B2, and he thinks it does more than they allow:  deep-sea montmorillonites "pick up even more sodium (plus other cations) in the base exchange sites than was present in the original igneous rock," which is his explanation for the excess of sodium over chloride in river water.⁶  Earlier he states the balance flatly:  "salt is recycling in steady state between the oceans and the continents ...  the same amount of salt leaves the oceans each year via ocean spray as is returned to the oceans via river waters."⁶

Cook's caveat should be stated too, because it is a real weakness in his position.  He does not quantify the clay sink.  "The quantitative answer to this question is much more difficult ... since there is incomplete information on clay compositions.  These are therefore omitted here."⁶  He asserts the sink rather than measuring it.  What he does instead is substitute a different clock:  uranium, which he thinks is not recycled, giving "an upper limit of about 10⁶ (10⁵ to 10⁷) years for the oceans."⁶  That is a million years, not ten thousand, and it fails by the same mechanism as the sodium version — reducing sediments and hydrothermal circulation are large uranium sinks, so "uranium is not precipitating from the oceans" is false.

Nobody worked the sodium budget before 1990.  The pre-1990 argument in the Creation Research Society Quarterly is Harold Camping's, and it is a different argument:  seawater is not saturated in most elements, and their residence times disagree by fourteen orders of magnitude, therefore the ocean is young.⁷  His three conclusions, in his own numbering:

"1. The oceans must be very young because small quantities of many of the elements are in solution.  2. The oceans must be very young because of the wide discrepancy of residency periods of various chemicals. ...  3. That salt (NaCl) and perhaps a number of other chemicals are in the oceans completely apart from normal rock weathering."⁷

The Creation Research Society's own commissioned survey of dating methods, in 1987, summarised the whole ocean argument in two sentences and named the same two authors:  "Camping (1974) has shown that most elements other than sodium which have dissolved in the sea give much shorter age estimates of the sea.  Morris (1975) reached similar conclusions."⁸  No budget, no sink inventory, no differential equation.  A full-text search of the CRSQ archive turns up no paper that performs a sodium input-output mass balance for the ocean, before or after 1990.⁹

That is a point in Austin and Humphreys' favour and this article states it plainly.  The 1990 paper was original work.  It was not a restatement of a movement talking point;  it replaced one, and it replaced it with something far more careful.

And the number they replaced was 4.2 times theirs.  Henry Morris's 1975 table of "uniformitarian estimates" gives "Influx of sodium to the ocean via rivers — 260,000,000 years."¹⁰  Austin and Humphreys give 62 million.  Two ICR-tradition publications, fifteen years apart, on the same quantity, differing by a factor of four, and no paper anywhere reconciles them.

Morris did not believe his own sodium figure either.  On the page after the table he writes that "the quarter-billion year age for the ocean determined by sodium influx neglects the certain fact (from both geological and paleontological evidence) that the ocean has been composed of salt water from the beginning."¹⁰  So the founder of the Institute for Creation Research set the sodium clock aside in 1975 on the initial-salinity objection, and ICR published it as its flagship ocean argument fifteen years later.

Austin and Humphreys, 1990

The paper is "The Sea's Missing Salt:  A Dilemma for Evolutionists," presented at the Second International Conference on Creationism in Pittsburgh and printed in the proceedings at pages 17 to 33.¹  It is free from the Institute for Creation Research, in a 2006 re-typesetting paginated 1 to 13.

Its abstract states the claim without hedging:

"The known and conjectured processes which deliver and remove dissolved sodium (Na+) to and from the ocean are inventoried.  Only 27% of the present Na+ delivered to the ocean can be accounted for by known removal processes.  This indicates that the Na+ concentration of the ocean is not today in 'steady state' as supposed by evolutionists, but is increasing with time. ...  A differential equation containing minimum input rates and maximum output rates allows a maximum age of the ocean of 62 million years to be calculated."¹

Two things about that framing deserve credit before anything else.  They do not claim to have dated the ocean, and they say so four separate times in the paper.  And they close by asking their critics for something specific and checkable:

"We challenge evolutionists and old-earth creationists to report quantitative data supporting a steady state ocean.  Those who propose that continental weathering and rivers have been delivering Na+ to the ocean for 3 billion years need to explain the sea's missing salt."¹

Thirty-six years later the honest answer to that challenge is that the budget does not balance, and that mainstream geochemists say so in print.  What follows is why that concession does not buy a young ocean.

The argument, worked in the open

The two tables

Everything turns on these.  Both are transcribed from the paper, in its own units of 10¹⁰ kilograms of sodium per year.  Ai is the present-day input, Aimin the smallest that input could plausibly have been in the past under the palaeo-scenario most generous to an old earth.

Sinks are 27% of sources.  That gap, 33.5 × 10¹⁰ kilograms of sodium a year, is the "missing salt" of the title.

The equations

The ocean's sodium inventory is the concentration times the mass of the sea:  10,760 mg/kg against 1.37 × 10²¹ kg gives Mp = 1.47 × 10¹⁹ kg.¹

Divide by the input and you get a filling time, τ = Mp/Ap = 32.2 million years.  Austin and Humphreys are careful about what that number is not:

"Evolutionists call τ the residence time of Na+, implying that τ is the average time a Na+ ion spends in the ocean.  However, as already demonstrated, the present oceans are not in steady state, so τ cannot be the residence time for Na+.  For clarity of concepts, we call τ the 'filling time'."¹

The filling time on its own is not a maximum age, because sinks remove sodium while the sea fills.  So they write each sink as proportional to the ocean's sodium content, Bi(t) = bi·M(t), and solve the resulting differential equation for the time to go from zero sodium to today's inventory:

T  =  -(τmax / xmax) · ln(1 - xmax)

where τmax = Mp/Amin and xmax = Bmax/Amin.  Feeding in the table totals gives τmax = 41.3 Myr, xmax = 0.58, and T ≤ 62 Myr.

A third number circulates, and it is not either of these.  Humphreys' own popular list of young-world evidences, and Andrew Snelling's Answers article built on it, lead with 42 million years:  "If seawater originally contained no sodium (salt) and the sodium accumulated at today's rates, then today's ocean saltiness would be reached in only 42 million years — only about 1/70 the three billion years evolutionists propose."¹¹  That is neither the 62 Myr bound nor the 32.2 Myr filling time.  It is a present-rates calculation with no differential equation and no palaeo-generous columns, and the closest thing to it that reproduces from Tables 1 and 2 is the inventory divided by the present net gain, 1.47 × 10¹⁹ ÷ 3.35 × 10¹¹ = 44 million years.  Running equation 8 on the present-day columns instead gives 37.5 million.⁵  Forty-two does not fall out of the paper's own tables by either route, and neither popular source shows its working.  The number the lay reader meets is not the number the peer-reviewed paper computed, and the popular sources present them as the same result.

It reproduces exactly

A Deno TypeScript script written for this article, research/salt-budget.ts, transcribes both tables and recomputes every published quantity.⁵  Nothing is fitted and nothing is tuned.

Every one of them lands.  No arithmetic error exists in this paper, and none of its published critics has found one either.  The dispute is entirely about the values in the two tables, not about what is done with them.  A reader who arrived expecting to catch a creationist in bad maths should stop looking and start reading the table entries, which is where the argument actually breaks.

One small internal inconsistency is worth a sentence, because it shows the tables were never fully reconciled with the running text.  The prose says "Addition of the calculated Na+ inputs from off-axial and axial hydrothermal springs gives the global hydrothermal vent flux:  A11 = 1.5 × 10¹⁰ kg/yr."  Table 1 tabulates 1.1, and the total of 45.7 is consistent with 1.1, not 1.5.¹  The difference is under 1% of the input and does not move the answer.

What the answer actually hangs on

Reproducing a calculation tells you nothing about whether it is robust.  The useful test is to move each input and see how far the answer moves with it.  The script does that one term at a time, dropping each input from both columns and rerunning equation 8.

Two things fall out of that table immediately.

Three of the eleven inputs contribute exactly nothing to the published bound.  Glacial silicates, glacier ice, and to a first approximation the dust and aerosol terms all have a minimum-past value of zero or near zero, and only the minimum-past column enters equation 8.  The paper advertises "seven previously unrecognized" sodium inputs.¹  Several of them do no work in the headline number.

The single most load-bearing number in the paper is the groundwater term, and it is the one number in it that rests on no measurement at all.

The groundwater input

Here is how A10 is derived, in full:

"The dissolved solids of ground waters can be measured in parts per thousand (ppt) and vary significantly from as fresh as rain water (0.0001 ppt) to extremely saline brine (over 250 ppt).  For this reason we have great difficulty in estimating their average Na+ content. ...  However, we know this average imperfectly.  If we assume that the average ground water has 26 mg/L of Na+ (5 times the concentration of average river water), the global Na+ flux is A10 = 9.6 × 10¹⁰ kg/yr from ground water seepage. ...  This rough estimate represents 50% of the total river input.  If further research should increase the global average Na+ content of the ground water, the flux of this very significant source would increase."¹

The water volume behind that number is defensible:  global rainfall minus evaporation runs about 10% higher than global river runoff, and they take the missing 3.7 × 10¹⁵ litres a year as groundwater seepage.¹  The concentration is not.  26 mg/L is a stipulation, five times average river water, chosen and stated as chosen.

To their credit, they flag it themselves.  What they do not do is show what happens if it is wrong.  Delete A10 and the ceiling moves from 62 million years to 109 million, a 77% change on a single unmeasured term.  Halve it and roughly half that shift appears.  No other term in the paper comes close.  An upper bound that moves this far on a stipulated concentration is not a bound anyone should quote to two significant figures.

The young-earth side has a live update to this term, and it is the water flux rather than the concentration.  Jonathan Sarfati reported in Creation in 1998 that radium measurements in coastal water had revised submarine groundwater discharge sharply upward:  "Previously, the amount of SGWD was thought to be a small fraction (0.01-10%) of the water from surface runoff, mainly rivers.  But this new study, measuring the radioactivity of radium in coastal water, shows that the amount of SGWD is as much as 40% of the river flow.  This means that the maximum possible age of the ocean is even smaller."¹²  The underlying paper is Moore's 1996 Nature letter on ²²⁶Ra enrichments in the South Atlantic Bight.  It does not repair the defect.  A larger water flux multiplied by a stipulated concentration is still a stipulated sodium flux, and the concentration is the term that carries the sensitivity.  It is fair to note it, and fair to say what it fixes and what it does not.

The mid-ocean-ridge sink

The other controlling number is on the output side, and it is the one place in this dispute where a live, current, mainstream disagreement exists.

Seawater circulates down through hot basalt at the mid-ocean ridges and out through the ridge flanks.  Sodium is exchanged with the rock on the way.  Austin and Humphreys put the whole ridge sodium sink at 0.44 × 10¹⁰ kg/yr for low-temperature basalt alteration and exactly zero for albitisation.  Glenn Morton, reading Holland's 1978 Table 5-11, reported 14 × 10¹⁰.  Morton's later calculation from vent-fluid chemistry gives 34.4 × 10¹⁰.  That is a factor of forty, on a quantity that is measurable in principle.

Sweep it and watch the bound:

At 16 units the maximum output-to-input ratio reaches one, the logarithm in equation 8 diverges, and nothing whatever can be said about the age of the ocean from sodium.  The entire result therefore turns on whether one number is nearer 0.4 or nearer 16.

That is not a rhetorical point.  It means the argument is not robust in the way an upper bound is supposed to be:  it does not degrade gracefully as the sinks grow, it disappears.  And Austin and Humphreys chose the low end of a forty-fold range on the strength of an argument about vent chemistry, over the printed warning of the author whose sink inventory they had adopted wholesale.

The four defects

One: the sink they set to zero is the one their own authority warns about

Austin and Humphreys take their sink list from Heinrich Holland's 1978 textbook The Chemistry of the Atmosphere and Oceans,¹³ and describe it as "those from the exhaustive list of Holland (1978, Table 5.14)."  Then they overrule its albitisation entry:

"The suggestion by Holland (1978, Table 5.14, p. 232) that albite formation is an effective sink for oceanic Na+ is not supported by the most recent data.  It appears that the seawater as it is heated from ocean temperature to 350 °C gains as much or more Na+ from low temperature solution of Na+ in basalt as is removed above 350 °C. ...  It is concluded that albite formation removes essentially no Na+ from the ocean.  Therefore, B6 = 0 kg/yr."¹

Page 232 is the page they cite, and page 232 says the opposite.  Holland's own summary of Table 5-14, in the paragraph that introduces it, singles out sodium:

"Sodium is a case in point.  Much of the river input of Na+ is certainly recycled as sea spray, and a good deal must be removed as a constituent of halite, but the removal rate via cation exchange on marine clays and by the formation of albite during the cycling of seawater through basalt near midocean ridges is probably not negligible."¹³

The same page also warns against reading Table 5-14 as a settled inventory at all:  "nearly all of the individual output rates are still only poorly known ... even today a rather cautious attitude toward proposed schemes of ocean outputs is advisable."¹³  Austin and Humphreys call the list "exhaustive."

Two hundred pages earlier in the same book, Holland tells his readers not to zero the albite term.  He is discussing his own Table 5-11, which tabulates what seawater gains and loses cycling through the ridges, and the sodium entry in it looks small:

"The apparently rather minor amount of Na removal in Table 5-11 is somewhat misleading.  The critical figure is really not the percentage of the total flux of river sodium removed, but the percentage removal of the river flux of the sodium derived from silicate weathering.  Since only about one-third of the total river flux has this origin (see Chap. IV), the removal of sodium during seawater cycling through midocean ridges is potentially of major importance."¹³

"Potentially of major importance" is the phrase, and it is about the term Austin and Humphreys set to zero.  The paper does not engage this passage.  Holland's point is exactly the one that matters:  most river sodium is recycled sea salt and dissolved old evaporite, so a sink that removes only a few percent of the total river flux may still be removing most of the genuinely new sodium.

Holland's reason for thinking the ridges matter is the heat-flow budget, and he states it with the uncertainty attached:

"The interpretation of the heat flow data is still uncertain, but it is hard to avoid the conclusion that the cycling of seawater through midocean ridges is a major mechanism for the removal of the river flux of Mg from the oceans unless the interpretation of the heat flow data turns out to be wrong."¹³

That is the honest form.  The magnesium case is strong;  the sodium case rides on the same water flux and is weaker, because sodium is exchanged in both directions.  Neither Austin and Humphreys nor Morton argues for the water flux they pick, and their two choices differ by a factor of ten.  Austin and Humphreys use Wolery and Sleep's stated upper limit of 2.3 × 10¹³ kg of seawater a year;  Morton uses the bottom of Holland's range, 2 × 10¹⁴.¹⁴  The number that decides this argument is a heat-flow number, not a sodium number, and neither side in the exchange goes and gets it.

The answers to defect one, and what they are worth

This is the article's strongest attack, so the replies to it belong on the same page.  There are two, both in print, and neither has been cited in the dispute as it usually runs.

Answer one:  Humphreys, quoted by Jonathan Sarfati in Creation magazine, December 1998.  Sarfati's article carries an addendum headed "is albitization the evolutionary escape hatch?", in which Humphreys is quoted verbatim with his permission:¹²

"What happens is this:  indeed albite forms in mid-ocean vents and takes sodium out of the high-temperature sea water.  But then when the albite gets into cooler water, it decomposes into the mineral chlorite and releases the same amount of sodium back into the sea water.  That is why albite (in any significant amounts) is found only at the mid-ocean ridges and nowhere else.  So his 'albite sink' would change into a 'chlorite source', and the net effect on sodium in the sea would be zero."

That is a mechanism, not a wave of the hand, and it deserves to be tested rather than dismissed.  It fails three ways.

Albite and chlorite are the same assemblage, not a reaction pair.  Holland's 1978 textbook — the book whose sink list Austin and Humphreys adopt — describes what altered ocean-floor basalt is actually made of:  "Typical greenschist assemblages in hydrothermally altered basalts dredged from the ocean floor consist of chlorite plus albite plus or minus actinolite plus or minus epidote."¹³  Chlorite sits beside albite in the same rock, not downstream of it.

The stoichiometry does not work.  Albite is NaAlSi₃O₈.  It contains no magnesium and no iron.  Chlorite is a magnesium-iron phyllosilicate.  Turning albite into chlorite is not a decomposition;  it is a replacement that requires importing magnesium and iron and exporting silicon and aluminium.  Humphreys names no reaction, no reference, and no rate.

The field claim is not right.  Albite is the sodium end-member of plagioclase and one of the commonest minerals in the crust;  read charitably as a claim about albitised basalt, it is still wrong, because albitised metabasalt is what obducted ophiolites are largely made of, and ophiolites are ocean crust that left a ridge and is still albitised.

Answer two:  Andrew Snelling, Answers magazine, 1 October 2012, in the article that is Answers in Genesis' flagship popular statement of this argument.  One sentence:¹¹

"Long-agers also argue that huge amounts of sodium are removed during the formation of basalts at mid-ocean ridges, but this ignores the fact that the sodium returns to the ocean as seafloor basalts move away from the ridges."

Two things about that sentence, and the second was checked in the page's raw HTML.

It contradicts the paper it defends, in the paper's own words.  Austin and Humphreys' term B5 is headed "Low Temperature Alteration of Sea Floor Basalt," and it is about exactly the rock Snelling says gives the sodium back:

"The coolest basalts which form the uppermost rock of the ocean floor also circulate seawater but the temperatures of these fluids usually remain below 60 °C.  At this lower temperature the basalt is weathered to form clay minerals. ...  The primary clay is saponite, a Na+-containing mineral of the smectite (montmorillonite) group.  Therefore, Na+ from low-temperature seawater reactions with the basalt must remove Na+ from seawater."¹

Cool sea-floor basalt is a sodium sink in their accounting, entered as a positive 0.44 × 10¹⁰ kg/yr in the removal column of Table 2.  If ageing crust returned sodium it would appear in Table 1 as an input, and it is not there.  Snelling's one-sentence rebuttal would require rewriting the table of the paper it cites.  Holland reports the same direction for the neighbouring alkali:  "K+ uptake during the weathering of basalt at the seafloor has been demonstrated, and may turn out to be an important K+ sink."¹³

Its footnote does not support it.  The clause stating the objection carries footnote 6, which cites Morton's 1996 letter — a fair and accurate attribution.  The clause stating the rebuttal carries footnote 7, which reads in full:  "Calculations based on many other seawater elements give much younger ages for the ocean, too.  See Stuart A. Nevins (Steven A. Austin), 'Evolution:  The Oceans Say No!' Impact no. 8 (Santee, CA:  Institute for Creation Research, 1973)."¹¹  That note says nothing about mid-ocean ridges, nothing about ageing crust, and nothing about sodium returning from basalt.  Nor does the article it points to:  Nevins 1973 is a sediment mass balance — 27.5 billion tons a year delivered against 820 million billion tons on the sea floor, giving 30 million years — and it contains no calculation from any dissolved seawater element at all.¹⁵  The one published sentence that answers this article's strongest defect carries a citation to an article about a different subject.

The fair verdict, stated in both directions.  The young-earth reply here is a mechanism with no magnitude attached, and its stated mineral reaction does not work as stated.  But this article's own criticism has the mirror-image weakness:  it can show that the ridge sodium sink is disputed by a factor of forty and that the bound vanishes above 16 units, and it cannot pin the number either, because the water flux is unmeasured on both sides.  Nobody in this dispute has gone and got the number.

Two: a quarter of the input comes from a table its own authors disown

A4, "Ocean floor sediments," is 11.5 × 10¹⁰ kg/yr, the second-largest input in the paper and 25% of the total.  It is sodium diffusing up out of sediment pore water into the sea, and the source is chapter 1 of Chemical Cycles in the Evolution of the Earth, by James Drever, Yuan-Hui Li and J. Barry Maynard.¹⁴  Austin and Humphreys give the derivation in one sentence:

"The estimate of Sayles (1979) is slightly adjusted for charge balance and yields A4 = 1.15 × 10¹¹ kg/yr (Drever, Li, & Maynard, 1988, Table 1.4, column 3)."¹

That is a correct reading of column 3.  Three pages before the table, the chapter's authors warn their readers about that very column:

"The fluxes usually conform to charge balance constraints at individual sites, but the averages do not.  Accordingly, the fluxes in column 3 of the table have been adjusted for electrical neutrality.  The values are high when compared with the net river flux.  For instance, note that the proposed diffusive flux would remove far more than the total available K+ and all of the Mg, whereas we believe that some Mg is used up in hydrothermal reactions at mid-ocean ridges."¹⁴

And two pages after it, having checked the pore-water estimate against an independent estimate from sediment compositions, they conclude:

"The Na+ is also very different, showing a large release to seawater in Table 1.4 but a large uptake into the sediments in Table 1.6.  Many of the analyses in Table 1.5 seem to include NaCl from sea salts that were not removed by washing the sediment.  Accordingly, the Na+ flux in Table 1.6 is almost certainly due to the burial of pore water, as in the calculation presented earlier, and not to diagenetic reactions.  In conclusion, it seems likely that the relative changes in Table 1.4 are correct but the absolute magnitudes are too high by a factor of at least 2."¹⁴

Read that carefully, because it says two separate things and both hit A4.  The whole table is at least twice too high by its authors' own estimate.  And the sodium entry in particular is the one that disagrees worst with the independent check, probably because the analysts did not wash the sea salt off the mud before measuring it.

Applying the correction the table's own authors ask for:

On its own that does not overturn the argument.  Combined with any non-trivial ridge sodium sink it does:  A4 halved plus a ridge sink of 10 units gives 126 Myr, and at 15 units the bound is gone.⁵

Three: an ellipsis in exactly the wrong place

This one is a matter of record and the reader should judge it for himself, so both halves are printed side by side.

Austin and Humphreys introduce the mainstream's episodic-evaporite answer in order to dismiss it:

"Evolutionists have claimed that the process of halite deposition (B4) is much different today than in the past.  They admit that modern marine halite deposits are of trivial volume, but attribute ancient massive halite deposits to short, irregularly occurring episodes.  Drever et al. (1988, p. 51) speak for many evolutionists who believe:  '. . . such events appear to be well able to absorb the river excess over long periods of time . . .'."¹

Here is that sentence whole, on the page they cite, with the next sentence after it:

"Therefore, such events appear to be well able to absorb the river excess over long periods of time, but their irregular occurrence suggests that the salinity of the oceans has varied substantially on a shorter time scale.  Holser et al. (1980) suggest that salinity may in fact have declined in spurts from about 45‰ in the Paleozoic to 35‰ in the modern oceans."¹⁴

The quotation is accurate as far as it runs and the page citation is correct.  This is not a fabricated quote and calling it one would be wrong.  What it is, is a cut made exactly where the sentence stops helping.

Two consequences follow, and the second is the serious one.

The paper misdescribes the position it is refuting.  The half-sentence quoted makes mainstream geochemistry look like it is gesturing at episodic events without evidence.  The half cut away states the mainstream position with numbers:  the ocean's salinity is not constant, it varies, and the direction of variation is down, from roughly 45 per mil in the Palaeozoic to 35 today.  Austin and Humphreys' abstract attacks "steady state ... as supposed by evolutionists."  The evolutionists they cite did not suppose it.

And a declining Phanerozoic salinity is the exact opposite of what their own model predicts.  Their creationist model, in their own words, "implies (a) that the initial level of Na+ in the ocean was a substantial fraction of today's level, (b) that there was a significant burst of input Na+ during the Genesis Flood, and (c) the Na+ input rate was at higher levels than today for thousands of years."¹  Every element of that produces a rising sodium curve through the rock record.  The sentence they cut says the curve falls.

Four: the world's rock salt is bigger than they were told

Halite deposition is the sink that would close the budget if it were large enough, and Austin and Humphreys cap it with a global inventory:

"The present inventory of rock salt in the earth's strata contains about 4.4 × 10¹⁸ kg of Na+ (Hosler's estimate of global salt inventory, as cited in Holland, 1984) which is 30% of the mass of Na+ in the ocean."¹

The name is William Holser, and the 30% figure is real:  Holland 1984 reports it as a personal communication of 1981.¹⁶  Spread over the Phanerozoic that inventory gives an average removal of 7.3 × 10⁹ kg Na/yr, and taking the Permian province alone and time-averaging it over the Permian gives their maximum, B4max = 4.0 × 10¹⁰ kg/yr.¹

Twenty-five years later the inventory had grown by a factor of two to three, mostly because of salt found under the deep ocean.  William Hay and five colleagues compiled the whole Phanerozoic evaporite record.³  They report:

"The present total inventory amounts to a minimum of 9.1 × 10⁶ km³ of halite, equal to about 19.6 × 10¹⁸ kg or more than 50% of the total of halite dissolved in the ocean today (36.8 × 10¹⁸ kg);  the maximum estimate is about 16.3 × 10⁶ km³ of halite, equal to 35.2 × 10¹⁸ kg, or 95% of the present total in the ocean."³

Sodium is 39.3% of halite by mass, so those figures are 7.7 × 10¹⁸ to 13.9 × 10¹⁸ kg of sodium:  1.75 to 3.15 times the 4.4 × 10¹⁸ kg Austin and Humphreys used.⁵  The reason is stated in the same paper.  Zharkov's compilation, which is Austin and Humphreys' source for the Permian share, is a compilation of Palaeozoic salt on land.  The Gulf of Mexico salt was only recognised after Sigsbee Knoll was drilled in 1967, and the deep-ocean Mesozoic salt of the Atlantic came later still.³

The charge is older than the paper.  Glenn Morton put it to a creationist audience at the First International Conference on Creationism in Pittsburgh in 1986, four years before "The Sea's Missing Salt" appeared, in a paper called "Geologic Challenges to a Young Earth."¹⁷  His target was Whitcomb and Morris's version of the salt clock in The Genesis Flood, and his argument is the one Hay et al. made quantitative twenty years later:

"This salt would have to be included in any calculation of the sodium age of the ocean because the quantity of salt in the oceans is only a part of the salt that has been carried into the ocean over time.  Is there any evidence of this being the case?  There is much salt that has been found contained within the sedimentary section.  A well drilled in northern Utah penetrated 7,600 feet of salt.  A well drilled off Nova Scotia drilled through 4900 feet of salt known as the Osprey Evaporites.  4,000,000 cubic kilometers of salt are found in the sediments below the Gulf of Mexico, and salt is found in the sediments in the bottom of the Mediterranean over an area of 1 million square kilometers and averages 1 kilometer in thickness."¹⁷

His conclusion is careful, and worth quoting because it is more modest than the use his rebuttal is usually put to:  "This does not necessarily mean that the oceans are old, it means we can't date them by sodium and chlorine content."¹⁷

The proceedings printed a formal discussion with four named respondents, and two of them were Steven Austin and John Morris.¹⁷  Austin's reply is a bibliography — four entries with parenthetical notes, no argument — and the only one that touches the salt is "D. I. Nutting, 1984, 'Origin of Bedded Salt Deposits:  A Critique of Evaporative Models and Defense of a Hydrothermal Model':  Unpublished M.S. thesis, Institute for Creation Research, 107 p."  John Morris's reply is one sentence:  "Evaporitic salt does contain abundant pollen grains.  However, the laterally extensive and thick salt beds found buried in the stratigraphic record do not, nor do they contain meteoric dust or wind-blown dust.  Morton's argument based on evaporitic salt seems to be a straw man."  Morton's printed closure:

"It is difficult to reply to a bibliography.  I do not claim to either have read everything nor can I cite everything I have read.  Dr. Austin does not raise any substantive issues but merely points out some of the literature that I didn't cite. ...  To say that salt does not contain pollen is to ignore the vast palynological literature on salt, some of which I cited in the paper.  Dr. Morris needs to read a few of them."¹⁷

That matters for more than chronology.  The largest sink in Austin and Humphreys' output table is halite deposition, and their cap on it rests on a global rock-salt inventory now known to be two to three times too small.  The objection that the inventory was too small was put to Austin in print, at the young-earth movement's own conference, before he published the paper, and it was answered with a reading list.  Nothing in the 1990 paper engages it, and a full-text sweep of all 457 ICC papers from 1986 to 2023 finds no later ICC paper that restates, defends or revises the sodium calculation.¹⁸

Morton also made the charge again in 1996 with an independent count, and it holds up.  From published volumes he computed 3.51 × 10¹⁸ kg of sodium in the Gulf of Mexico and central North Atlantic salt alone, plus 3.1 × 10¹⁷ kg in the Zechstein, giving 3.8 × 10¹⁸ kg of sodium in three deposits, "86% of what you say is in the whole world."¹⁹  He also reports that Austin, in a letter of 10 May 1996, allowed "that the world-wide resource of sodium was 1.47 × 10¹⁹ kg in rock salt," a figure 3.3 times the published one and at the top of Hay's range.¹⁹  That concession is attested only by Morton's report of private correspondence and no published creationist statement of a revised figure has been found, so it should not be treated as established.  The point does not need it.  Hay et al. is peer-reviewed, and it says the same thing.

Scaling B4max by the factor Hay's inventory implies pushes the bound to 68 to 87 million years by itself.⁵

There is a further asymmetry worth naming, because it is structural rather than numerical.  A3, "Rivers:  chloride solution," 7.5 × 10¹⁰ kg Na/yr, is mostly sodium released by dissolving old evaporites.  That is not new sodium entering the system.  It is ocean sodium coming back, the return leg of a loop whose forward leg is B4.  Austin and Humphreys count the return leg at its full present rate and cap the forward leg with a long-run average.  Hay et al. hindcast the return leg directly and get 13 to 42 × 10⁹ kg of chloride a year over the Phanerozoic, 29.6 to 34.6 × 10⁹ for the Quaternary and Holocene, which as sodium is roughly 2 × 10¹⁰ kg/yr rather than 7.5.³  Substituting their figure moves the ceiling to 79-83 million years;  simply closing the loop, by setting B4max equal to A3, gives 69.⁵

Hay et al. are also blunt about the number Austin and Humphreys' A3 depends on.  It comes from Meybeck's global average river chloride, and they write:

"Meybeck's (1979) global average for [Cl−] in rivers is dominated by those rivers which drain areas with large evaporite deposits in the continental interior.  The total discharge of the rivers for which he had reliable data is only about 60% of the global discharge from land to sea, and the assumption that Cl− has the same average concentration in the other 40% of the world's rivers, which drain more coastal areas mostly lacking evaporite deposits, may not be justified."³

What the calculation assumes, stated precisely

Equation 8 returns a maximum age under three conditions.  It is worth taking them one at a time, because two of the three are commonly attacked and only one of them actually fails.

One:  the ocean began with zero sodium.  This is not a defect in their argument.  They set the initial sodium to zero deliberately, "to make our age estimate as large as possible for the benefit of the evolutionary model."  For an upper bound that is correct reasoning:  any non-zero starting value makes the derived time shorter, not longer.  A critic who attacks the zero-initial assumption has misread the paper.  What the assumption does destroy is any reading of 62 million years as an estimate rather than a ceiling, and Austin and Humphreys say so themselves.

Two:  the system is closed and monotonic, so sodium only accumulates and every sink varies smoothly with how much sodium is in the sea.  This is the real failure, and it is falsified by a measurement the paper itself supplies.  Equation 3 writes each sink as Bi(t) = bi·M(t), proportional to the ocean's sodium content.  Evaporite deposition does not work that way.  It is switched on and off by basin geometry, tectonics and climate, and it can run at rates no smooth function of ocean chemistry would predict.  Austin and Humphreys concede this in a footnote to their own reference list:

"If the 'Messinian salinity crisis' is assumed to have had a duration of one million years, B4max would be 5.8 × 10¹¹ kg/yr, a value slightly greater than all the combined inputs.  The 'Messinian salinity crisis,' however, is admitted by many to be a truly extraordinary event.  It cannot be used to estimate the long term removal rate of Na+ in halite."¹

Their refusal to average the Messinian into a long-term rate is defensible.  Averaging an extraordinary event into a mean rate would be bad practice.  But look at what the footnote concedes.  By their own arithmetic, one documented evaporite episode removed sodium faster than every input in their Table 1 combined.  An event that happened cannot be excluded from a budget on the ground that it was unusual.  Once a single measured episode can outrun the entire input, the assumption that carries equation 3 is dead, and no amount of care with the table values revives it.

Three:  the input and output terms are right.  Half-failed, as the four defects above show.  Two of the four largest inputs are unsound as stated:  groundwater rests on a stipulation, and the pore-water diffusive flux comes from a table its own authors halve.  Together those are 46% of the total input.  On the output side the one term set to zero is the one Holland warned not to dismiss, and the halite cap uses an inventory now known to be two to three times too small.

What survives, and it is not nothing.  A genuinely unclosed sodium budget does exist in the mainstream literature.  Drever, Li and Maynard's own Table 1.8 carries a river excess of 2.2 × 10¹² moles a year, which as sodium is about 5 × 10¹⁰ kg/yr.¹⁴  That is real, admitted, and roughly a sixth the size of the imbalance Austin and Humphreys report.  The difference between the two figures is almost entirely their eight extra input terms, of which the two largest are the two the mainstream literature either does not accept or explicitly downgrades.  They close their excess with episodic evaporites and they show the arithmetic:  one Miocene event absorbs eleven million years of it.  That is an answer, not an evasion.  But it is an answer that requires the ocean's salinity to move around rather than sit still.  So the honest statement is that the sodium budget is not in steady state, mainstream geochemistry has said so in print since at least 1988, and the mainstream resolution is a fluctuating ocean.  Austin and Humphreys are right that steady state is false.  They are wrong that not-steady-state means monotonically rising.

The measured record of seawater

Everything so far is a dispute about fluxes, and fluxes are hard.  There is a way to bypass the whole argument:  measure ancient seawater directly.

When seawater evaporates and halite crystallises out, tiny pockets of the brine get trapped inside the growing crystals.  In "chevron" halite the inclusions line up in bands parallel to the growth faces, which is the textural signature of crystals that grew on the floor of an evaporating brine body rather than being recrystallised later.  Those pockets are samples of the surface brine at the moment of crystallisation, and with a scanning electron microscope and an energy-dispersive x-ray detector you can freeze them and analyse the major ions in place.

Tim Lowenstein and four colleagues did this across the Phanerozoic and published the result in Science in 2001.²⁰  Their abstract:

"Systematic changes in the chemistry of evaporated seawater contained in primary fluid inclusions in marine halites indicate that seawater chemistry has fluctuated during the Phanerozoic. ...  The data suggest that seawater had high Mg2+/Ca2+ ratios (>2.5) and relatively high Na+ concentrations during the Late Precambrian [544 to 543 million years ago (Ma)], Permian (258 to 251 Ma), and Tertiary through the present (40 to 0 Ma), when aragonite and MgSO4 salts were the dominant marine precipitates.  Conversely, seawater had low Mg2+/Ca2+ ratios (<2.3) and relatively low Na+ concentrations during the Cambrian (540 to 520 Ma), Silurian (440 to 418 Ma), and Cretaceous (124 to 94 Ma), when calcite was the dominant nonskeletal carbonate."²⁰

Read the order.  Sodium high in the Permian, low in the Cretaceous, high again now.  Not a ramp.  An oscillation, with a period of 100 to 200 million years, and one full cycle running the wrong direction inside the sequence Austin and Humphreys' model needs to be climbing.

Two things make this hard to wave away.  The first is that the sodium signal is the one that survives the obvious objection.  A critic might say the inclusion chemistry was altered by dolomitisation of the surrounding carbonate platforms, which would change Mg and Ca.  Lowenstein and colleagues answer that directly:  "the systematic changes in the Na+ concentration of fluid inclusions (Fig. 1) cannot be explained by dolomitization."²⁰  The second is that the oscillation is not one measurement.  It moves in step with the mineralogy of marine limestones, the mineralogy of potash evaporites, sea level, and inferred sea-floor spreading rate, four independent records built by different people for different reasons.²¹

The volume record says the same thing by a completely different route.  Hay and colleagues reconstructed ocean salinity from the compiled volumes of evaporite deposits, using chloride as the tracer because chloride is the one ion essentially confined to the ocean.  Their conclusion:

"The data suggest that there have been significant changes in the mean salinity of the ocean during the Phanerozoic. ...  Palaeozoic salinities were in the low to high 40's and reached the 50's (‰) in the Early Cambrian, depending on the evaporite data."³

Today's ocean is 34.7 per mil.  Their curve runs downward, and the largest steps down are the great salt extractions into the young ocean basins that opened as Pangaea broke up.  Holland's earlier and more conservative bracketing, in his 1984 sequel, gives the same sign¹⁶:  the total halite inventory limits how much higher salinity could ever have been, and bromide in basal halites limits how much lower, so Phanerozoic seawater sits between about 35 and 45 per mil.^{16 16}

Here is the honest counter a young-earth reader will make, and it deserves an answer rather than a brush-off.  If the whole Phanerozoic column was laid down in a single year of Flood, then "600 million years of oscillating salinity" is not 600 million years of anything.  It is one year.  A record that took a year to form cannot refute a young ocean by its duration.  That objection is correct on its own terms, and the salinity record does not, by itself, refute Flood geology.

What it does refute is the slow accumulation model that this argument's own arithmetic assumes.  And there is a second problem the compression does not remove.  The fluid-inclusion record is not a flat line with noise on it.  It is a structured signal in which brine chemistry, the mineralogy of the enclosing limestones, and the mineralogy of the potash salts all change together, twice, in phase.²¹  A single year of Flood deposition has to reproduce two complete chemical cycles in a stratigraphic sequence laid down in months, with the trapped brine at every level matching the carbonate mineralogy of the rock around it.  That is a chemistry problem, not a dating problem, and no published Flood model has attempted it.

Holland states the limitation of the method fairly and it should be repeated here:  halite is highly soluble and therefore prone to recrystallisation, so the pedigree of any given inclusion is a live question.  "It should, however, be possible to find brine inclusions whose pedigree is above suspicion."¹⁶  Lowenstein's answer to that is petrographic:  chevron texture, and the fact that inclusions from geographically separate basins of the same age overlap each other while none of them overlaps the modern seawater evaporation path.²⁰

The argument most readers have met is a different argument

Everything above is about Austin and Humphreys' sodium budget.  Most people who have encountered "not enough salt in the sea" have not met that argument.  They have met a different one, and it fails for a different reason.

Morris, 1974: shortest residence time wins

Scientific Creationism is the best-selling young-earth textbook ever printed.  Its ocean section runs three pages and its headline number is not sodium:²²

"Entirely apart from the meteoritic nickel question, the mere fact that the ocean's nickel content could have accumulated from river inflow in about 9000 years seems to set an upper limit to the age of the ocean, unless it can be shown that this dissolved nickel either precipitates out on the ocean bottom, or is somewhere returned through the atmosphere to the continents.  Neither has been shown."²²

The method is then generalised to every element:

"The same kind of calculation can be made for other dissolved chemicals in the ocean.  That is, the amount of any given chemical in the ocean, divided by the annual increment of that chemical through river inflow, will yield the time required to accumulate the chemical, assuming none was present in the ocean to begin with and the inflow rate has always been the same."²²

He then prints a table of results taken from Riley and Skirrow's Chemical Oceanography.  Here are its ocean rows as reprinted in his 1975 CRSQ version, which was read for this article rather than reconstructed:¹⁰

That table refutes the argument it is printed to support, and it does so in one line.  If each entry were an upper limit on the age of the ocean, the entries would have to agree.  They span six orders of magnitude.  Lead says two thousand years, sodium says two hundred and sixty million, and both are computed the same way from the same book.  At most one of them can be an age.

The reason they disagree is not mysterious.  A stock divided by an influx is a residence time, and a residence time is short precisely when the sink is fast.  Silicon's residence time is short because diatoms build shells out of it.  Lead's is short because it is scavenged onto sinking particles.  Aluminium's is a hundred years because it is insoluble and falls out.  Every one of those sinks was known when Morris wrote, and the book his table comes from — Chemical Oceanography — is a book about the sinks.  Taking residence times out of a treatise on removal mechanisms and then writing "Neither has been shown" is the defect in the popular argument, and it is checkable inside Morris's own footnote.

Morris half-saw it, and that should be said.  Two pages after the table in the 1975 version he writes that "the age of the ocean as based on the influx of lead and aluminum is considerably less than man's own recorded history of the ocean.  It must be that the rates of influx of those metals were much lower in the past, or else that they are continually being removed from solution in some way."¹⁰  So he grants a sink for lead and aluminium in order to discard the short numbers, and refuses one for nickel in order to keep the short number he wants.  The rule that selects which entries count is not stated anywhere.

Austin and Humphreys do not make this mistake.  Their whole argument is that the sodium imbalance is 0.27, not that sodium's residence time is short, and equation 8 correctly returns "less than infinity" when the imbalance vanishes.  They picked the element with the longest residence time and the best-inventoried sinks, which is the opposite of what Morris did.  Grading the two arguments by the same standard means saying that the technical version is much better than the popular one, and that the popular one is the one in circulation.

The same method, run on nickel again in 2016

Creation Ministries International revived Morris's nickel calculation in Creation magazine in 2016 and got the same kind of answer:  nickel at 0.228-0.693 parts per billion in seawater against 1-3 ppb in fresh water gives "an age of the oceans is 8,200 years," and against the United Kingdom's 30 ppb environmental quality limit, "this limit would be reached with current nickel input in just 1,076,000 years."²³

To its credit, the article does go after the sink.  Manganese nodules contain 1.25 to 1.5% nickel and there are about 500 billion tonnes of them, which absorbs "168,000 years maximum" of river input;  subduction is dismissed because the sea floor being subducted is no older than 200 million years.²³

The nodules are the wrong sink to test.  Nickel's residence time in seawater is of order 10⁴ years because it is scavenged onto sinking organic and mineral particles and buried in the ordinary sediment column, which is a far larger reservoir than the nodule field armouring its surface.  An inventory of the armour does not bound the flux through the floor beneath it.

And a reader who has just met that argument will ask whether the nodules are old at all, so the answer belongs in the same paragraph.  Kenneth Patrick argued in the Journal of Creation in 2010 that manganese nodules are themselves young:  they are found almost entirely at the sediment-water interface rather than through the column, and observed growth rates are orders of magnitude above the 1-10 mm per million years textbooks give.²⁴  His fast rates are real measurements, and they come from a 1930s artificial reservoir in the Altai at 1.7-1.8 mm a year, from Oneida Lake in New York, and from nodules grown on Second World War shell splinters at 0.6-1 mm a year — shallow, oxygenated, high-flux settings with a manganese supply per unit area nothing like the abyssal Pacific.  Patrick says so himself, and it is the answer to his own paper:

"it just won't do to apply, say, those rates of >1 mm/yr directly to MNs growing at the ocean seafloor.  To do so would see MNs grow to diameters of greater than 3 kilometres in only a few million years!"²⁴

Read the paper closely and the whole growth-rate case rests on transferring those rates to the abyss, which its author has just said cannot be done.  It leaves the mainstream measurement untouched:  deep-sea nodule growth rates are obtained radiometrically from profiles inside individual nodules, not by dividing a size by an assumed age, and Patrick nowhere attacks that measurement — he calls it "radiometric analysis" and moves past it.  His second argument, that nodules cluster at the sediment surface rather than running through the column, has a standard physical answer he does not address:  nodules are kept exposed by burrowing animals and bottom currents that roll them, which is why they are the one sea-floor object that does not get buried.  His own source reports that once a nodule is buried it "neither grows nor dissolves,"²⁴ which is what that answer predicts.

Grading the rebuttals by the same standard

The standard rebuttal contains a category error

The version of the answer most readers meet is the TalkOrigins Index to Creationist Claims, entry CD221.1, four sentences long, relying entirely on Glenn Morton:

"A detailed analysis of sodium shows that 35.6 × 10¹⁰ kg/yr come into the ocean, and 38.1 × 10¹⁰ kg/yr are removed (Morton 1996).  Within measurement error, the amount of sodium added matches the amount removed."²⁵

Those two numbers are not comparable, and the entry does not say so.  35.6 is Austin and Humphreys' Amin, the minimum past input under the palaeo-scenario most generous to an old earth, the column in which glacial silicates and glacier ice are set to zero.  38.1 is Morton's revised maximum past output column.  Austin and Humphreys' present input is 45.7.  Morton's own letter prints his input list at the top, and it is the Aimin column copied verbatim:  5.0 for river sea spray, 6.2 for ocean-floor sediments, 0 for glacial silicates, 0 for glacier ice, 9.3 for groundwater.¹⁹

Recomputed against the right column:

The invented term is "collective small outputs," 3.6 × 10¹⁰ kg/yr, 9.4% of his total.  It is not a measurement.  His stated rule is that each of the sixteen dashes in Holland's 1978 output table represents a process running at 1.25% of the river flux.¹⁹  His point that a dash in a table does not mean zero is fair.  Turning sixteen dashes into a number and adding it to a budget is not.  Strip it out and his total falls to 34.5, below even the figure he was balancing against.⁵

And his headline sink, "Alteration of Basalt 14 × 10¹⁰ kg/yr," is attributed to a table in Holland 1978 that could not be read for this article.  Printed page 199 is a full-page table on a print-disabled archive.org scan;  the search-inside index carries no paragraph text for that leaf and the page images return 403.  It has not been verified and it has not been refuted.  Holland's own commentary on the facing page calls the table's sodium removal "apparently rather minor," and fourteen units would be 73% of the total river sodium flux, which nobody would call minor.  So either Morton read a different column, or a different unit, or Holland's "minor" refers to something else.  Until that page is read, 14 × 10¹⁰ is Morton's claim about Holland, not Holland's figure.

The Berean Archive's position is that the most widely circulated rebuttal of this argument is wrong, and that saying so is not a favour to creationism.  It is what grading both sides by the same standard requires.  The entry has stood since 2004.

Where Morton is right

Two of his substantive points land, and one of them is the strongest thing anybody has written against the 1990 paper.

The bedded salt deposits.  He quotes their paragraph arguing that modern marine halite accumulation is trivial, then asks:

"Where in your discussion here do you include the Mediterranean salts, the Zechstein salt of Germany, the Louann Salt of the Gulf of Mexico, the Osprey Salt of Offshore Canada, or the Salina salt of New York?  All of these are bedded in the middle of the geologic column and represent huge episodic removals of salt from the oceans by evaporation.  No one believes that similar deposits are being laid down today.  These occurrences are episodic, not continuous."¹⁹

That is exactly right, it is checkable, and Hay et al. 2006 confirmed the quantitative version of it fifteen years ago.  It also has priority:  he had made the same argument in the 1986 ICC proceedings, and Austin had answered it there with a bibliography.¹⁷

The albitisation calculation.  Morton's second letter, from November 1997, is much better than his first and is almost never cited.²⁶  He goes to Von Damm's compilation of vent-fluid analyses, takes twenty-six vents, computes a mean sodium-to-chloride molar ratio of 0.800 against seawater's 0.8545, converts the deficit to 29.7 millimoles of sodium removed per kilogram of circulated seawater, multiplies by a seawater cycling rate of 2 × 10¹⁴ kg/yr, and gets 3.44 × 10¹¹ kg of sodium a year.  That single term exceeds the entire missing 3.35 × 10¹¹.

This is the right kind of argument:  a measurement-based estimate of the one term Austin and Humphreys set to zero.  Two problems keep it from being decisive as printed, and both should be stated.

The water flux is disputed by an order of magnitude.  Morton uses 2 × 10¹⁴ kg/yr, the bottom of Holland's 1978 range.  Austin and Humphreys use Wolery and Sleep's 2.3 × 10¹³ kg/yr, which is stated in the source as an upper limit.¹⁴  Scale Morton's calculation by that figure and his sink falls to about 4 × 10¹⁰, which does not close anything.  Neither writer argues for his choice.

The sampling has a physical problem Morton does not address.  Look at his own table:  vent chlorinities run from 176 to 1990 millimoles per kilogram against seawater's roughly 545.  That spread is produced by sub-sea-floor phase separation into a low-chlorinity vapour and a high-chlorinity brine.  The brine stays in the crust and never reaches a sampler.  A sodium deficit measured in the vapour-enriched fluid that does reach the sea floor therefore cannot be integrated straight into a net oceanic sodium loss, because the complementary phase is missing from the sample set.  Von Damm herself draws only a qualitative conclusion from the deficit, that it "provides our best evidence that albitization is an active process in hydrothermal vent systems."²⁶  The global flux is Morton's step, not hers, and Von Damm's chapter was not obtained for this article, so both of his quotations from it rest on his transcription.

The criticism nobody cites: Menninga at the conference

The ICC proceedings printed a formal discussion alongside the paper, with the authors' closure.  The discussant was Clarence Menninga of Calvin College,²⁷ a co-author of Science Held Hostage, which Austin and Humphreys had attacked by name.  This is old-earth Christian criticism, published beside the argument in 1990, and almost nobody quotes it.

Menninga's core objection, repeated in four of his seven points:

"However, even if the residence time of sodium in the oceans is less than the 260 million years reported in much of the recent literature, we must remember that the residence time of sodium in the oceans is not the same as the age of the earth.  After all, the residence time of aluminum in the oceans is only 100 years, and that is not viewed as a dilemma."²⁷

"Sodium is not the only element with a residence time in the oceans which is longer than several thousand years.  The residence time for potassium is 11 million years, for magnesium is 45 million years, for silver is 2 million years, and for uranium is 500,000 years.  Are all of those long residence times in error?"²⁷

Those are Henry Morris's own numbers, thrown back at his institute.  Potassium 11,000,000, magnesium 45,000,000, silver 2,100,000 and sodium 260,000,000 are four rows of the table Morris printed in CRSQ in 1975, and the fifth, uranium at 500,000, is the figure Morris attributes to Riley and Skirrow on page 154 of Scientific Creationism.^{22 10}  Menninga was not importing an outside objection.  He was reading the Institute for Creation Research's own most-reproduced table back to the Institute for Creation Research, and asking why the same arithmetic that was proof in 1974 was an error in 1990.  Neither the discussion nor the closure mentions where the numbers came from.

And his philosophical jab, which is the sharpest thing in his piece:

"After gathering a considerable amount of data, and after performing several calculations and logical analyses, the authors suggest that the scientific study they have done isn't worth anything, after all.  They suggest that God might have made the oceans recently, with a great deal of sodium (and other elements) already dissolved in the water.  If that is the attitude one wishes to adopt, what is the justification for doing the scientific study?  A paper which is only one or two sentences in length would suffice to reach the same result."²⁷

Austin and Humphreys win this exchange on points, and this article says so.  Their reply to the aluminium objection is correct and worth quoting in full because it is the clearest statement in the whole paper of what they think they have shown:

"Aluminum's small residence time is not a dilemma for old-earthers.  Hence, he implies sodium's residence time should not present a dilemma, either.  But it is not the residence time which makes the dilemma;  it is the imbalance between sodium inputs and outputs.  We can see this by contrasting what Equation 8 says about aluminum and sodium.  The data for aluminum gives xmax = 1;  using this in Equation 8 tells us that the age of the ocean is equal to or less than infinity.  For the data we report concerning sodium, Equation 8 tells us that the ocean is less than 62 million years old.  Both statements are true, but the one based on sodium is more stringent."²⁷

That is a good answer.  A residence time is a ratio of stock to flux and says nothing about age.  An input-output imbalance is a different quantity, and equation 8 is a statement about the imbalance.  Menninga's objections 2 through 5 all attack a claim the paper does not make, and the paper had already said in its own text that the filling time "is not the age of the ocean."  He also supplied none of the quantitative data he was asked for, and Humphreys names that:  "his review contains no such data."²⁷

But the closure concedes something without noticing.  Answering Menninga's sixth point, they say the study "rigorously tests a hypothesis."  Their own alternative model has three free parameters:  initial sodium, the size of the Flood surge, and the duration of elevated post-Flood input.  None of the three has an independent constraint.  A model that can absorb any measured sodium inventory by adjusting its starting value is not tested by the sodium inventory.  Menninga's jab lands harder than the reply allows.

Objections and answers

"The sinks you invoke to close the budget are unquantified hand-waving.  Name one, with a number, that is actually measured."

This objection is fair against most published rebuttals and it should not be dodged.  Morton's "collective small outputs" is an assumption with a rule attached.  His 14 × 10¹⁰ basalt figure has not been read off the page it is credited to.  His 34.4 × 10¹⁰ albitisation flux is built on a water flux ten times what the other side uses, from a sample set that systematically misses one of two phases.  None of that is measurement in the sense the objection demands.

What is measured is the halite inventory, and it moves the answer in the same direction.  Hay et al. counted the salt:  19.6 to 35.2 × 10¹⁸ kilograms of halite, which is 7.7 to 13.9 × 10¹⁸ kilograms of sodium, two to three times the 4.4 × 10¹⁸ Austin and Humphreys were told, with the increase coming mostly from deposits found under the deep ocean after 1967.³  That is a count of rock, not a model.  It does not close the budget on its own.  It does show that the sink most likely to close it was underestimated by a documented factor.

"The world's rock salt never came out of the sea.  It is juvenile brine degassed from the mantle along deep faults.  So the halite inventory is not sodium the ocean lost, and the fluid inclusions in it are not samples of seawater.  Two of your four defects and your whole seawater record go with it."

This is the strongest objection on this page, it comes from inside the young-earth movement, and it has to be answered rather than waved at.  Henry Morris and John Morris state it in Science, Scripture, and the Young Earth, following the Ukrainian geologist V. I. Sozansky:²⁸

"Evaporites are even more clearly the result of rapid processes, in spite of their very misleading name.  The typical uniformitarian explanation, repeated by Young, is that these great thicknesses of salt beds, gypsum beds, and the like have been formed by slow and cyclically repeated evaporation from inland lakes or relict seas.  Such an explanation, however, is increasingly recognized today as impossible, mainly because the 'dirty' salt deposits forming today are quite unlike those found in the geologic record, remarkable in their purity."²⁸

"Salt is not an evaporitic formation or a derivative from volcanic rock;  it is a product of degasification of the earth's interior.  The salt precipitated from juvenile hot water which emerged along deep faults into a basin as a result of change in thermodynamic conditions."²⁸

Granted, that would be fatal to two sections above.  The halite inventory would not be a sink at all, and Lowenstein's fluid inclusions would be trapped hydrothermal fluid rather than trapped seawater.  It is still a live young-earth position:  Stef Heerema, Gert-Jan van Heugten and Timothy Clarey, ICR's staff geologist, presented a salt-magma origin for the Permian Castile Formation at the Ninth International Conference on Creationism in 2023, and named Morton's 1986 ICC paper as the challenge they were answering.²⁹

Four answers, in ascending order of force.

One:  the same book endorses the opposite.  Forty-seven pages after the juvenile-brine passage, the same authors write:  "An important young age argument has been developed by Dr. Austin and Dr. Russell Humphreys. ...  it can be shown that all the sodium in the ocean can be accounted for in only 62 million years."²⁸  Austin and Humphreys' largest possible sink is halite deposition, B4max = 4.0 × 10¹⁰ kg/yr, and it is computed by treating the world's rock salt as sodium that left the sea.  Both claims cannot be true.  The young-earth literature has not settled what evaporites are, while resting its flagship ocean calculation on a specific answer to that question.

Two:  the objection deletes a quarter of the argument's own input.  A3, "Rivers:  chloride solution," 7.5 × 10¹⁰ kg Na/yr, is mostly sodium redissolved from bedded evaporites.  If the salt was never marine, that sodium is not returning to the sea from the sea;  it is new sodium arriving from the mantle by a route with no known rate.  Delete A3 from Austin and Humphreys' table and the ceiling moves from 62 million years to 94.5 million.⁵  Taking the salt magma seriously makes the bound weaker, not stronger.

Three:  Sozansky's own stated reason was refuted at a creationist conference in 1986.  The reason Morris quotes is "the absence of remains of marine organisms in ancient salts."²⁸  Glenn Morton's ICC 1986 paper is about exactly that, with references:  bedded halite carries fossil pollen grains, and bacteria and plant spores "in quantities from 50 to 2,000 spores per cubic centimeter," and the salt is conformably interbedded with ordinary sandstone and shale and has bedding of its own — "if it were metamorphic, one would not expect the salt to be so conformable to the surrounding sediments."¹⁷  John Morris replied that the thick buried beds contain no pollen;  Morton's closure was that this "is to ignore the vast palynological literature on salt."¹⁷  Neither side has published a measurement since.

Four:  the halite carries a marine chemical fingerprint, in two independent tracers.  Bromine substitutes for chlorine in halite with a distribution coefficient well below one, so the bromide content of the lowest salt in a basin records the Br/Cl ratio of the water that fed it.  Holland compiles the measurements and they fall in the band expected from seawater:  "Bromide in the lowermost salt sections of various basins, as a measure of the Br/Cl ratio of ancient seawater ...  The upper section indicates the range of values expected from modern seawater."¹⁶  Holland states the caveat honestly — the data scatter, and "conclusions based on this criterion alone must be treated with caution" — so it is the second tracer that carries the weight.  The fluid inclusions do not merely contain brine;  they lie on evaporation paths, curves in Na-Cl and Mg-Cl space whose shape is set by which salts crystallise out and when, with a kink at the point where halite starts precipitating.  Lowenstein's team verified the method on present-day halite first:  "Fluid inclusions from present-day marine halites overlap the modern seawater evaporation paths, indicating that fluid inclusions from modern halites contain unaltered evaporated seawater brines."²⁰  The ancient inclusions lie on paths of the same form, displaced;  and the displacement moves in step with the mineralogy of the limestones around them and of the potash salts above them.²¹  A brine rising along a fault has no reason to track the carbonate mineralogy of the contemporaneous sea.

What the objection does survive with.  Neither Sozansky's model nor the Castile poster has been tested against the bromine profiles or the inclusion evaporation paths by anybody on either side, and this article cannot cite a paper that has done it.  The answer above is an argument from evidence already collected for other purposes, not a study aimed at the salt-magma claim.  That study has not been done.

"Residence time is not age.  Dividing a stock by a flux tells you how long the stock would take to fill, not how long it has been there.  The whole thing is a category error."

This objection is not fair to Austin and Humphreys, and the article says so plainly.  It is the objection Menninga made in 1990 and they answered it correctly.  Their argument is not "the residence time of sodium is 32 million years, therefore the ocean is 32 million years old."  Their argument is that the ratio of outputs to inputs is 0.27, which if true means the sea is filling, and that a filling reservoir with a known present contents and a bounded fill rate has a computable maximum age.  Equation 8 is a statement about the imbalance, not about the residence time, and it correctly returns "less than infinity" when the imbalance vanishes.

Anyone who uses this objection is attacking a claim the paper explicitly disclaims, in a paper that says four times over that it has computed an upper limit and not an age.  The argument fails on its inputs, not on its logic.

"The ocean could simply have started out salty.  Halley said so in 1715."

He did, and he drew the same conclusion Austin and Humphreys draw:  a salty start makes the derived date a maximum.  So this objection does not touch the bound.  It has teeth in only one direction, and it is aimed at the creationist model rather than at the old-earth one.  The Austin-Humphreys model requires the ocean to have been created with sodium "a substantial fraction of today's level," topped up by a Flood surge, then topped up further for thousands of years at elevated rates.  Three free parameters, no independent constraint on any of them, and a prediction of rising salinity that the halite inclusions contradict.  Halley's objection is a problem for the young-earth model, not a rescue of the old-earth one.

"Creationists never engage the fluid-inclusion record, because it is the one measurement that settles this."

Largely true as a matter of citation record, and the omission matters.  Lowenstein's 2001 paper is the direct measurement of ancient seawater sodium, and the young-earth literature on ocean salinity does not discuss it.  The 1990 paper predates it, which is a complete excuse for the 1990 paper and no excuse at all for the popular restatements that have run since.

But the engagement, when it comes, has a real move available and this article should name it rather than pretend otherwise.  If the entire Phanerozoic column was deposited in one year, the inclusion record is a year's worth of chemistry and its duration proves nothing.  What the record still demands, and what no Flood model has attempted, is a mechanism that produces two full in-phase cycles of brine chemistry, limestone mineralogy and potash mineralogy in a sequence laid down over months.  There is a second move available, and it is stronger:  deny that the halite is marine at all, which is the salt-magma objection answered above.  Neither move has been worked against the inclusion data by anybody on either side.  That is the question worth putting, and it is a chemistry question.

"62 million years is 6,000 times more than a young earth needs.  Why would a creationist cite this at all?"

Because the claim is comparative, not absolute.  Austin and Humphreys are not arguing that the ocean is 62 million years old;  they say explicitly that their result "leaves room for the possibility that the sea is less than 10,000 years old."  The argument's whole force is that it is fifty times short of what mainstream geology needs.  Judge it on that, which is what this article has done.

The precedent: the movement has already retracted this argument's twin

Creation Ministries International maintains a page called "Arguments we think creationists should NOT use," which exists so that "arguments that we think should no longer be used" can be named in public.³⁰  None of the ocean-chemistry arguments is on it.  One structurally identical argument is:

"For a long time, creationists claimed that the dust layer on the moon was too thin if dust had truly been falling on it for billions of years.  They based this claim on early estimates — by evolutionists — of the influx of moon dust, and worries that the moon landers would sink into this dust layer.  But these early estimates were wrong, and by the time of the Apollo landings, NASA was not worried about sinking.  So the dust layer thickness can't be used as proof of a young moon (or of an old one either)."³⁰

Look at the shape of that.  An inventory, divided by a present-day influx, read as a maximum age.  The influx estimate came from the mainstream literature the creationists were quoting.  It turned out to be wrong, and the argument was dropped.  That is exactly the shape of the sodium argument, and exactly the reason its halite cap fails:  the global rock-salt inventory Austin and Humphreys were given, on a personal communication of 1981, is now known from a peer-reviewed compilation to be one-half to one-third of the real figure.

This is the fairest thing that can be said about the movement's standard here, and it cuts both ways.  It shows the young-earth movement does retract on this argument-form when the number moves, which is more than most of its critics allow.  It also shows that the number has moved on sodium and the retraction has not followed.  The last word from the Institute for Creation Research, in Acts & Facts in November 2021, restates Austin and Humphreys unchanged — eleven sources, seven sinks, 62 million years — and adds that "when more realistic numbers were used within a creation model, such as a nonzero initial concentration of Na+ and the likely effect of the Genesis Flood most likely adding salt to the oceans, a biblical time frame of thousands of years was achieved."³¹

Where the young-earth case is strongest

Three things, none of them softened.

The arithmetic is clean and the claim is stated honestly.  Every published result reproduces.  They said four separate times that they had computed an upper limit rather than an age.  They named their weakest assumption in the text rather than burying it.  They asked their critics for a specific, checkable thing.  That is better practice than most of what gets published on either side of this argument.

Their reply to Menninga is correct, and it is still correct.  The commonest answer to this argument, "residence time is not age," misses the target.  Thirty-six years on, that answer is still the one most sceptics reach for, and it is still wrong.

The budget really does not balance, and the mainstream literature says so.  Drever, Li and Maynard carry a river excess of about 5 × 10¹⁰ kg of sodium a year and close it with episodic events that, by construction, are not happening now.  Hay and colleagues close their 2006 paper by asking whether the deep-sea salt deposits are as large as they think and whether there is "an unknown mechanism for regulating the salinity of the ocean."³  The geochemistry here is not tidy, and a rebuttal that presents it as tidy is misleading its readers.

Where it fails

The two numbers that carry the result are the two weakest numbers in the paper.  Groundwater is a stipulation and the answer moves 77% on it.  The ridge sodium sink is disputed by a factor of forty and at the high end the bound does not exist.  An upper bound that evaporates when one contested term reaches a value inside its published range is not a bound.

Three of the four largest terms have documented problems, and two of the three come from the paper's own sources.  The pore-water input is taken from a table its authors halve.  The halite cap uses an inventory since revised upward by a factor of two to three.  The zeroed albitisation sink is the one Holland warned, in the book they call exhaustive, is "potentially of major importance."

The differential equation cannot represent what the rocks record.  Sinks that vary smoothly with ocean chemistry cannot produce the Messinian, and the Messinian is in the paper's own footnote, running faster than every input combined.

And the direction is wrong.  The model predicts sodium rising through the record.  Fluid inclusions in halite show it oscillating, high in the Permian and low in the Cretaceous.  Evaporite volumes show mean salinity declining from the low 50s per mil in the Early Cambrian to 34.7 today.  A young-earth reader can answer that the whole record is one year long, and that answer is available.  It is not an answer to the arithmetic in this paper, which assumes slow accumulation from the start.

The verdict.  The sea's missing salt is real.  It is not evidence for a young ocean.

Six items bearing on this topic could not be verified against the source they are attributed to, including the Holland 1978 table that carries Morton's largest sink.  They are listed with what was tried in quarantine, and none of them is used above.

Sources