Comments on "A Geological Time Scale 1989" by Harland et al

Comments on "A Geological Time Scale 1989" by Harland et al

I do set My bow in the cloud

After a number of proddings by correspondents having an evolutionary viewpoint, I finally obtained a copy of (Harland, W.B.; R.L. Armstrong; A.V. Cox; L.E. Craig; A.G. Smith; and D.G. Smith, 1990, "A Geologic Time Scale 1989," Cambridge University Press, Cambridge). This book is the standard reference work for assigning ages to periods of the geological column. Here are my impressions of it.

Does Harland et al Support the Validity of Isotopic Dating?

I expected to find something in this book to justify the geological time scale, but I was disappointed. The book (Dalrymple, G.B., 1991, "The Age of the Earth," Stanford University Press, Stanford, CA.) does a much better job justifying the age of meteorites (assuming a constant decay rate) than Harland et al (1990) does in justifying radiometric dating on the Phanerozoic (Cambrian and later strata). Dalrymple (1991) shows that multiple isochrons on the same meteorite give consistent dates, which is hard to explain by any other mechanism than a lapse of time. But Harland et al (1990) simply gives a table of dates and the parent-daughter combination used to obtain them, without mentioning whether concordant dates were obtained with different methods, or whether isochrons were used, or whether the isochrons used were based on different minerals or not. Thus Harland et al gives no way to assess the quality of the dates used. Multiple mineral isochrons are particularly reliable since they can hardly be explained by mixings, and concordances of different methods are also convincing, though they can be accidental. From a brief scan of parts of the table of dates in Harland et al, there do not seem to be many concordances of different methods on the same formation. There was not even a table in Harland et al giving the number or percentages of dates obtained by different methods in their list of dates.

Another serious problem is the small number of dates cited in Harland (probably substantially under 800 dates in all). This may seem like a large number, but not when one considers that over 100,000 dates have been measured altogether. It is true that the selected dates show a reasonably consistent pattern of increasing age with the conventionally accepted order of geological periods, and that different methods on a given period give fairly concordant results among the selected dates. But the method of selecting these 800 or less dates was not based on how reliable any single date was, but rather on how well they agreed with the main body of dates:

The criteria for exclusion include rejection by the original authors, excessive uncertaintly in date or stratigraphic position ... . Items that are clearly anomalous with respect to the main body of data have also been excluded.

Furthermore, even the selected dates often have quite a bit of scatter, as may be seen from the charts on pp. 119-137.

Their table (pp. 82 - 101) has an overwhelming preponderance of K-Ar dates, with a few Ar-Ar dates that are measuring nearly the same thing. Thus any pattern in the data, if it exists, can be explained by some property of K-Ar dating, and not necessarily by an agreement between different methods. A few more dates are fission track dates, and since fission track dating is calibrated in terms of other methods, this does not add any evidence for the correctness of the dates. The relatively small number of other dates that agree with the K-Ar dates might be explained by pure coincidence. It is noteworthy, for example, that some geological periods contain many more Rb-Sr or U-Pb dates than others, suggesting that the Rb-Sr and U-Pb and other dates are only concordant with K-Ar dates on certain geological periods. Particularly for dating of glauconies, which form the majority of the dates for many geological periods, K-Ar dates are often given but Rb-Sr dates comparatively infrequently. Since glauconies can be dated by both methods, this also suggests that the dates are discordant. It would be helpful to see the Rb-Sr dates of all glauconies whose K-Ar dates are given, and vice versa, to see if there really is some agreement between the methods. Even if the two methods are generally within 10 or 20 percent of one another, this would at least be a pattern needing explanation.

There is a list of geological formations given, in an accepted order of age, and so it would be interesting to see the pattern of all dating methods on these formations. Let's call these formations F1,F2, ..., Fn, listed in their accepted order of age. Let D1,D2, ..., Dn be the distributions of all dates obtained on these respective formations by some method, such as U-Pb dating of zircons, or Rb-Sr dating, or whatever. Then let these distributions of dates be published, so we can see if there is some consistent pattern which needs explanation, at least for the Phanerozoic. If other dating methods are all roughly correlated with the K-Ar dates on these formations, this is evidence that these dates indicate a passage of time, relative to a constant decay rate. Otherwise, the discordances suggest that the different methods do not agree, and argues against the reliability of radiometric dating on the Phanerozoic.

The Role of Evolution in Constructing the Geological Column

On page 4, Harland et al (1990) state:

Biologic evolutionary history, especially for Phanerozoic time, has given us not only the principal means of time-correlation but the basis of the unique progressive traditional stratigraphic scale.

The problem with this is that evolution is used to construct the geological sequence, and the geological sequence is used to prove evolution.

Hints of Trouble with K-Ar Dating

I found the following two statements about K-Ar dating to be bombshells:

Decades of experience have shown that concordant and stratigraphically consistent dates can be obtained using hornblende, biotite, and high-temperature feldspar of lavas and high-temperature pyroclastic rocks ... . Low K minerals are less suitable because of larger analytical errors and greater vulnerability to contamination, or to problems with excess Ar.

Of course, it is disturbing that "decades of experience" are cited rather than some a priori reason for believing these minerals to be reliable. The reason that these minerals are preferred is that they give the expected results. But the potentially most damaging comment is that minerals low in K tend to have more excess Ar, and thus give K-Ar dates that are too old. This suggests that much or all of the argon in minerals has nothing to do with the potassium, but entered by some other means than radioactive decay. Thus minerals with less potassium tend to have larger K-Ar ages, and are viewed as having greater problems with excess argon.

The second quotation is the following:

As a rule, the amount of atmospheric Ar in a whole-rock sample increases with age and amount of alteration ... . This effect is notable even before the sample is visually suspect.

This is saying that older rocks have more atmospheric argon. So there is some means by which, over long time periods, more and more atmospheric argon is absorbed into a rock. It is also reasonable to assume that the same is occurring with argon diffusing up from the mantle. This latter argon is nearly indistinguishable from radiogenic argon because of its very low argon 36 content. This will cause the K-Ar dates of rocks to increase with time, simply because they have more time to absorb argon! This could easily explain the pattern of K-Ar dates. I also recall that argon absorbed in this way can become firmly embedded within the rock, and thus impossible to detect as excess argon. Dickin (Radiogenic Isotope Geology, 1995, p. 247) mentions a study showing that volcanic rocks contain excess atmospheric argon, some of which cannot be removed by baking in a vacuum. Faure (1986, p. 74) also states concerning removing atmospheric argon before dating, "The difficulty can be reduced, if not completely avoided, by the removal of adsorbed atmospheric argon before the argon is extracted from the samples. It has been claimed that this can be accomplished by preheating samples under vacuum or by leaching them briefly with hydroflouric acid, or both ... . However Armstrong (1978) has questioned whether atmospheric argon, that has been acquired by minerals over a long interval of time, can be removed by this method." This further substantiates the suspicion that much or all of the argon measured in K-Ar dating is argon absorbed after the formation of the rocks.

Problems with Various Dating Methods

Chapter 4 of Harland et al is enlightening because of its frank discussion of the problems of various dating methods. For example, concerning K-Ar dating of glauconies, we read (page 75-76)

with care in the selection of well-formed, high K glauconies, the K-Ar dates obtained are in good agreement with high-temperature mineral dates. But the material is very susceptible to Ar loss due to thermal or tectonic overprints ... and there are cases where a detrital component leads to apparent Ar excess, especially in young samples ... . Given the many reservations expresssed about this mineral in the past, its relatively weak Ar retentivity, and the fluid-rich chemical environment of sedimentary rocks, this observation [discordant dates] is no suprise.

These observations are significant, since K-Ar dates of glauconies constitute a large fraction of the total dates in the given table of dates.

On page 76 we read

Incipiant weathering can affect Rb-Sr dates of biotite - resetting them to meaningless lower values ... - so that concordance of both Rb-Sr and K-Ar is important.

This quotation suggests that sometimes the Rb-Sr dates are concordant on biotite and sometimes they are discordant.

On page 77 we read

Glaucony may be dated by Rb-Sr but rarely with the precision available using K-Ar methods ... . Rb-Sr dates for glauconies provide a consistency check, but are usually less definitive, and are about as easily reset as K-Ar dates. Initial ratio uncertainty and recycled glaucony are problems that must be addressed.

Again, we have that sometimes Rb-Sr and K-Ar are concordant and sometimes not. The "initial ratio uncertainty" means uncertainty as to the amount of parent and daughter product initially in the sample.

On page 78 we read

The precision and accuracy of fission-track dates is a thoroughly debated topic ... . The accuracy is dependent on calibration using fission-track standards separated from rocks that have been dated by other techniques.

Since fission track dates are calibrated using other methods, it is not surprising that the dates obtained should often be concordant with these same other methods.

Zircons and the concordia-discordia method for dating them are emphasized in theoretical treatments of dating methods. However, on page 77 we read,

The real-world problem is that zircon suites with only an age of crystallization and one later event are rare and that the analyses are tedious and expensive, so that only a few points of time scale significance have been published.

They also mention that young zircons (up to 100 Ma of assumed age) are difficult to date using discordia because the concordia and discordia lines are so close. Older zircons, however, are prone to radiation damage and gain or loss of lead. A number of other problems with zircons are cited. They mention (p. 78) that U-Pb and Pb/Pb dates are rarely the same, and list some of the "tricks" that have been "invented" to make zircons concordant.

On the positive side, a study is listed (p. 79) in which Sm-Nd and U-Pb dates agreed to better than one percent. It would be interesting to know if this formation overlies fossils.

The Branching Ratio

There is a discussion of the use of decay constants beginning on page 190. It appears that nearly the correct branching ratio and other decay constants for K-Ar dating have been agreed upon since 1976, and began to appear in dates published about 1978. However, nearly the same values of constants for K-Ar dating have been used since the mid fifties (p. 76).


Though it is well written and informative, Harland et al (1990) does not give evidence of the correctness of the geological time scale, and even gives hints of problems with isotopic dating. Such a justification was apparently not the purpose of this work. If not, then some other study dealing with this issue is urgently needed.

Back to home page.