Source: “Dinosaurs – An Adventist View” David C. Read (reposted with permission)

Radiometric dating, sometimes called isotopic dating, is an attempt to determine when igneous rocks were formed based upon the rocks’ content of certain chemical elements. The technique is based on the principle that an unstable, radioactive parent element “decays” into a stable daughter element at a known average rate. For example, potassium-40 changes into the gas argon-40 at the rate of one half, every 1,280 million years.1 This means that after 1,280 m.y., half of the potassium-40 has changed into argon-40, and after 2,560 m.y., half of the remainder, or three-quarters of the original, has changed into argon, and so on.2

Proponents of radiometric dating argue that by measuring the ratio of parent element to daughter element and then applying the rate of decay, one can calculate how long the parent has been decaying into the daughter. For example, if one can determine the relative amounts of potassium-40 and Argon-40 and then apply the 1,280 m.y. half-life figure, one should be able to calculate how long the radioactive potassium has been decaying into argon, which, in turn, is supposed to indicate how much time has passed since the rock was formed. It’s a beautiful theory but, like so many beautiful theories, it does not work in reality.

Radiometric dating is based upon several assumptions, all of which must be true for the method to work. First, the average rate of decay must be known and must have remained constant since the rock was formed. Second, the proportion of daughter product in the rocks at the time the rocks were formed must be known. If there was unaccounted-for daughter product in the rock to begin with, the rocks will have a built-in appearance of age.

Third, the rock must be a “closed system.” An hourglass provides a useful example of a closed system. No sand gets into or out of an hourglass; the sand merely falls from the top to the bottom. The same must be true of the parent and daughter elements in tested rocks, if they are to keep time like an hourglass. None of the parent or daughter elements must have been migrating into or out of the rock sample tested. The isotopic system must have been as sealed off as an hourglass.

It is easy to appreciate why, if the method is to work, these assumptions must reflect reality. If sand has been escaping from the bottom of an hourglass, an observer noticing the small amount of sand in the bottom will conclude that the sand has not been running as long as it really has been. By contrast, if sand has been added to the bottom of an hourglass from some source other than the top, an observer would be fooled into believing that the sand had been running much longer than it actually had. If the isotopic system is an open system, radiometric dating will not work.

Astonishingly, dating scientists freely admit that the rocks are open systems. In fact, the openness of the rocks is one of the most frequently employed explanations for the many “wrong” dates yielded by the isotopes. It is claimed that the rocks lose or retained daughter product for any number of reasons, including recrystallization, leaching, heating, permeability, depth of burial, grain morphology, weathering, secondary alteration, and metamorphism.3

Bad dates are very common, but there is no way to know how common, because most bad dates are never published in scientific journals. They are just tossed into the laboratory wastebasket without any explanation. “If all the rejected dates were retrieved from the waste basket and added to the published dates,” argues Richard Milton, “the combined results would show that the dates produced are the scatter that one would expect by chance alone.”4

Since many radiometric dates are wrong—possibly most of them, who knows?—and it is acknowledged that most rocks are open systems and therefore unsuitable for radiometric dating, it is vital that there be some way to tell, prior to initial radiometric testing, whether a given rock sample has become “open.” Otherwise, dates can be accepted or rejected on a post hoc basis depending upon whether they fit with preexisting expectations. Those dates that seemed right based upon preexisting expectations would be accepted, and those that seemed wrong would be rejected. In other words, unless there is some way to tell, prior to initial testing, whether a rock sample is suitable for testing, the entire “science” of radiometric dating could be based on circular reasoning.

There are numerous admissions in the literature, however, that it is not possible to determine, prior to preliminary dating, whether a given rock sample has become an open system. If the date derived is within the expected range of results, the rock is assumed to have been a closed system. If the date is “wrong,” it is assumed that the rock was not a closed system, and one of the standard post hoc rationalizations—such as “cooling age,” reheating, leaching, inherited argon, argon loss (or other daughter product loss), metamorphism, etc.—is employed to explained the “bad” results.

There is a fascinating story that illustrates many of the problems with radiometric dating. It concerns attempts to date a layer of rock in Kenya. Known as the “KBS tuff controversy,” this episode involved several of the leading authorities on radiometric dating.5 The scientists involved became antagonistic toward each other, and almost everyone in anthropology during the 1970s took one side or the other. It is, as one Darwinist commentator wrote, “a story that demonstrates how very unscientific the process of scientific inquiry sometimes can be.”6

Anthropologists have long combed through the Lake Turkana (formerly known as Lake Rudolf) area in Kenya, searching for the putative missing links between apes and humans. In 1969, Richard Leakey, son of Louis and Mary Leakey and a leading anthropologist in his own right, had set up camp on the shore of the lake. Kay Behrensmeyer, then a graduate student in paleontology at Harvard and a member of Leakey’s team, found several primitive stone tools embedded in a layer of “tuff,” which is rock formed from volcanic ash. This particular rock layer became known as the KBS (Kay Behrensmeyer Site) tuff.

Leakey supplied samples from the KBS tuff to Jack Miller, a geophysicist at Cambridge University, to determine the feasibility of radiometrically dating it. Jack Miller and his partner Frank Fitch preliminarily dated the rocks at 212 to 230 million years old.7 In conventional geochronology, those dates correspond to the Triassic (the lower dinosaur strata), but Leakey was working near the boundary of the Tertiary and the Quaternary, which is thought to be only around two million years old. “From these results,” wrote Fitch and Miller, “it was clear that an extraneous argon age discrepancy was present.”8

Fitch and Miller believed that the tuff, instead of being deposited when ash settled out of the air, had been formed when rivers or streams had washed the ash down from highlands into a floodplain. They hypothesized that during this process the ash had picked up minerals from older deposits. But they did not identify any exposed Triassic rock that could have contributed material to the KBS tuff, which suggests that this was a post hoc rationalization concocted after the initial results missed the anticipated “ballpark” by two orders of magnitude.

Fitch and Miller sent back for more samples from the KBS tuff. They tested these and reported a preliminary age of 2.4 million years. They told Leakey that there were two techniques available to give a more exact radiometric age: 1) the traditional potassium argon technique and 2) a newer, more sophisticated technique called argon-40 argon-39. This latter technique involves converting the radiogenic potassium to the gas argon-39, then simultaneously measuring both types of argon in a machine called a mass spectrometer. This method was about twice as expensive as the older method, but it could work with smaller samples and would yield a much more exact measurement. The newer technique, they told Leakey, “would result in the tuff being incontrovertibly dated and with greater accuracy than any other site in Africa or elsewhere.”9 Leakey chose the technique that promised the “incontrovertible” date. Fitch and Miller ran the tests and reported that the tuff was 2.61 million years old, plus or minus 0.26 million years.

That result was within the “ballpark,” and thus not very controversial.10 A few years later, however, it was to become extremely controversial. In 1972, one of Leakey’s African associates, Bernard Ngeneo, found a hominid skull below the KBS tuff. This was skull KNM-ER 1470. As we discussed in Chapter 17, Darwinists can point to relatively few fossils as possible transitional forms between apes and humans. As a result, individual fossils can loom large in attempts to construct an evolutionary lineage or “family tree.” Individual hominid fossils can, and often do, change the whole evolutionary picture.11

Such was the case with skull 1470. Although it had a very small cranial capacity, in general morphology (at least as dubiously reconstructed by Leakey12) it resembled a modern human skull much more closely than did the Australopithecines (“southern apes”). If the KBS tuff was 2.6 million years old, then skull 1470, which was found below the tuff, must be even older than that—probably almost three million years old. But the Australopithecines had been dated at about 1.9 million years old—one million years younger than the much more human looking skull 1470.

The relative ages gave the impression that, over the course of a million years, a more human looking primate had evolved into a more ape-like primate. Evolution was going in the wrong direction! Dated as it was, skull 1470 exploded the then-current theory about how and when humans had evolved. “Either we toss out this skull,” said Leakey, “or we toss out our theories of early man.”13 Anthropologists were not about to toss out their theories of early man. Something had to give.


Figure 1: KNM-ER 1470. KNM-ER stands for Kenya National Museum—East Rudolph

It did. Garniss Curtis, of the University of California at Berkeley, was called upon to re-date the (already incontrovertibly dated) KBS tuff. He dated samples at 1.6 million years old and 1.8 million years old, about a million years younger than Fitch and Miller’s date. Curtis’ result was much more in line with the Darwinian assumption that fossil hominid crania should be moving from a more ape-like to a more human-like condition, not the other way around. Had Curtis been diplomatic toward Fitch and Miller, the controversy might have ended there, but, adding insult to injury, Curtis criticized Fitch and Miller’s samples, their methodology, and their laboratory techniques. The gauntlet had been thrown down.

Taking up the gauntlet, Fitch and Miller redid their tests, using a new and supposedly more accurate rate of decay, and found that the KBS tuff was 2.42 m.y. old. Most tellingly, Fitch published the fact that Curtis had gotten a “scatter” of results ranging from 1.5 to 6.9 million years. How had Curtis settled on 1.6 million years when he could have selected 6.9 million years? Obviously, 1.6 million coincided with previous expectations, and 6.9 million did not.
Fitch reported that his scatter of results had only ranged from 0.5 to 2.64 million years, implying that his technique was better because his scatter wasn’t as large. In understated but deliciously snide language, Fitch wrote of Curtis:

. . . K-Ar apparent ages in the range 1.6-1.8 Myr obtained from the KBS Tuff by other workers are regarded as discrepant, and may have been obtained from samples affected by argon loss.14

Fitch’s suggestion that Curtis’ samples were affected by argon loss illustrates the ease with which the open system excuse is invoked after the fact to explain away unwanted results.

In 1980, Ian McDougall of Australian National University weighed in on the KBS tuff controversy, reporting a date of 1.88 million years. Again, much more interestingly, McDougall revealed that Fitch and Miller had achieved a scatter of 0.52 to 2.64 million years on one set of samples, but on another set, they had gotten a scatter of 8.43 to 17.5 million years! Thus, the total scatter was from half a million years to 17.5 million years, a range of 17 million years (not including the preliminary finding of 230 million years) on a layer of rock supposedly around two million years old. Said Garniss Curtis, of Fitch and Miller’s work: “Their choice of 2.6+ for the age of the KBS Tuff seemed to be by their reaching into a hat filled with all the numbers they had obtained and coming out with 2.6 m.y.”15

The enormous range of results recorded by the competing groups of dating scientists illustrates Richard Milton’s point that if all results were published, not just those in the expected ballpark, the combined results would show the scatter that could be expected from random chance.

Meanwhile, Ian McDougall’s group had some scatter of its own to deal with, as one of their samples gave an age of 4.11 million years and another sample gave an age of 7.46 million years. These dates were of course rejected:

We attribute these poorly reproducible ages to the presence of variable but small amounts of old detrital K-feldspar in the aliquants used in the argon extractions. Careful petrographic examination of the mineral concentrate, however, did not lead to positive identification of detrital K-feldspar. Nevertheless, there is no doubt that old detrital material was being brought into the East Turkana Basin during deposition of the sediments.16

In other words, McDougall believed that his bad dates were a result of older material being included in the tuff when it was deposited, or re-deposited (which, recall, was the same story Fitch and Miller came up with to explain their 230-million-year date).

McDougall’s group could find no trace of this putative older mineral material, even after very carefully examining the rock, but they know it was there because . . . well, because the samples gave bad dates. This illustrates that there is no way to tell, by examining a sample, whether it is suitable for radiometric dating. If the result is in the expected ballpark, everything is assumed to be okay. If not, then the sample must have been unsuitable, even though a careful examination revealed no problems.

The controversy was ultimately settled at around 1.8 million years, not on the basis of radiometric dating—any date between 500,000 and 17.5 million years could have been selected with equal justification based purely on the isotopes—but on the basis of pig fossils.17 Here, as in every other case where the radiometric dates have conflicted with fossil evidence, fossil evidence prevailed, and conflicting radiometric dates were ignored or explained away.18 But the faunal remains were not clear, either. Correlation of elephant fossils seemed to support Fitch and Miller’s date; some researchers claimed that the pig fossils supported a younger date, while others claimed that the pig fossils supported Fitch and Miller’s date.19

But neither the faunal remains nor the isotopic results were really driving the KBS Tuff debate. Rather, it was the mess that skull 1470 made of the human evolution story that created the pressure to re-date the Tuff. Before the problems created by skull 1470 became too urgent, everything seemed to support Fitch and Miller’s date. Faunal correlation, paleomagnetism, and fission-track dating all impressively and concertedly supported their date.20 After it became clear what mischief the initial dating of the KBS tuff would do to the human evolution story, everything changed. It turned out that everything was negotiable, except the need to come up with a plausible presentation of the putative apes-to-humans transition.

Because this controversy was fought in print, the public learned how many different dates were tossed into the wastebasket. Several factors came together, including: 1) the discovery of a fossil that looked like it might fit into the evolutionary story of human origins; 2) the relevant layer of rock had already been dated before the fossil was found; 3) the dating of this fossil was crucial because, in terms of evolutionary theory, the fossil seemed to be out of sequence; 4) three different groups of dating scientists worked on the problem; and 5) these groups were not diplomatic in commenting on each other’s work.

Were it not for this combination of circumstances, we would never have known how many, and how widely varying, were the discarded dates, not to mention the facile post hoc explanations for them. Regarding this episode, Roger Lewin wrote, “[T]here is a degree of uncertainly in science that is not often made public, because it is contrary to the mythology of what science is supposed to be like.”21 The KBS tuff controversy shows that the reality of science resembles politics—maybe even theology!—more closely than it resembles its own carefully nurtured public image of detachment and objectivity.22

By now it should be obvious that radiometric dating is not what we have been led to believe. It is nothing like an exact science. Even after taking pains to test only samples that will yield a preconceived result, there is typically a large scatter of results for the dating scientist to choose from.23 Based upon prior expectations, which, in turn, are based upon geology, stratigraphy, index fossils, professional consensus, and what seems reasonable—none of which have anything to do with the isotopes per se—one of the dates is selected as “accurate.” After the winning candidate is chosen, the other dates are ignored or explained away by appealing to one (or more) of a staggering array of excuses. The post-hoc rationalization chosen depends only by the dating scientist’s imagination. The entire process is an exercise in prejudice confirmation. As a method of getting at truth, it is utterly worthless.

1. Roth, Ariel A. Origins: Linking Science and Scripture (Hagerstown, MD: Review and Herald, 1998) p. 251.
2. Actually, most of the K-40, about 89 percent of it, degrades into Ca-40; only 11 percent degrades into Ar-40.
3. Woodmorappe, Studies in Flood Geology, pp. 158-160.
4. Milton, at p. 51.
5. See, generally, Lubenow, Marvin L., Bones of Contention (Grand Rapids, MI: Baker Book House, 1992), pp. 247-266, appendix entitled, “The Dating Game.” See also Roger Lewin, Bones of Contention (New York: Simon and Schuster, 1987), pp. 189-252; Richard Milton, supra, at pp. 53-56; Michael A. Cremo, Richard L. Thompson, Forbidden Archeology: The Hidden History of the Human Race (Los Angeles: Bhaktivedanta Book Publishing, Inc., 1993), pp. 693-702. F.J. Fitch and J.A. Miller, “Radioisotopic Age Determinations of Lake Rudolf Artifact Site,” Nature, 226:226-228 (April 18, 1970).
6. Lewin, Roger, Bones of Contention (New York: Simon and Schuster, 1987), p. 190.
7. Lewin, at pp. 192, 194, 195; Lubenow, at p. 249.
8. F.J. Fitch and J.A. Miller, “Radioisotopic Age Determinations of Lake Rudolf Artifact Site,” Nature, 226:226-228 (April 18, 1970).
9. Lewin, at p. 192 (emphasis added).
10. There was one note of doubt, however. During the late Pliocene/early Pleistocene, pig molars were increasing in length and height. Basil Cooke, an expert on fossil pig teeth, noted that the pig teeth found in the area of the KBS tuff seemed to indicate that the tuff should be dated younger than 2.6 million years. Lewin, at pp. 195-197.
11. In Ever Since Darwin, Gould wrote that each year when the topic of human evolution “comes up in my courses, I simply open my old folder and dump the contents into the nearest circular file. And here we go again.” In his next book, The Panda’s Thumb, he writes, “I’m mighty glad I wrote [those words] because I now want to invoke that passage to recant an argument made later in the same article.” See Gould, Ever Since Darwin (New York: W.W. Norton & Co., 1977), p. 56; The Panda’s Thumb (New York: W.W. Norton & Co., 1980), p. 125. Another example concerns Neanderthal Man. In the Flamingo’s Smile (New York: W.W. Norton, 1985), p. 41, Gould wrote that Neanderthal was “a race of our own species, not an ancestor or any form of ‘missing link’.” But in Leonardo’s Mountain of Clams and the Diet of Worms (New York: Harmony Books, 1998), p. 210, Gould wrote that the Earth of 40,000 years ago contained three coexisting human species Homo neanderthalensis, Homo erectus, and Homo sapiens. Evidently he had come to believe, in the meantime, that Neanderthal Man was a separate species. Clearly, human evolution is not a subject where there is a lot of stability or continuity of concepts.
12. In 2007, Dr. Timothy Bromage announced a reconstruction of KNM-ER 1470 that was considerably more apelike that Leakey’s reconstruction. “Dr. Bromage”s findings call into question the extent to which H. rudolfensis differed from earlier, more apelike hominid species. . . . “Dr. Leakey produced a biased reconstruction based on erroneous preconceived expectations of early human appearance that violated principles of craniofacial development,” said Dr. Bromage, whose reconstruction, by contrast, shows a sharply protruding jaw and a brain less than half the size of a modern human”s.
13. Lubenow, at p. 253, quoting Richard Leakey, “Skull 1470” National Geographic, June 1973: 819.
14. Fitch, Hooker, and Miller, “Argon 40/Argon 39 dating of the KBS Tuff in Koobi Fora Formation, East Rudolf, Kenya,” Nature 263:741 (October 28, 1976, italics added) as quoted in Lubenow, p. 256. Later, Fitch also suggested “overprinting” as an explanation for the younger dates. Lewin, at p. 203.
15. Lewin, at p. 233, quoting Garniss Curtiss.
16. McDougall, Maier, Sutherland-Hawkes, and Gleadow, “K-Ar age estimate for the KBS tuff, East Turkana, Kenya,” Nature 284:230-31 (March 20, 1980) as quoted in Lubenow, p. 263 (emphasis added).
17. Lubenow, at pp. 255-266 (“The pigs won. . . . The pigs won over the elephants. The pigs won over K-Ar dating. The pigs won over Ar-40 Ar-39 dating. The pigs won over fission-track dating. They won over paleomagnetism. The pigs took it all.”); Lewin, pp. 195-252. Regarding scatter, Lewin writes that “at a conference in Nairobi held in September 1973 [Fitch and Miller] presented 41 separate age determinations on the KBS tuff, which varied between 223 million and 0.91 million years.” Lewin, at pp. 194, 195
18. Differences between radiometric data and fossil data “are a usual phenomenon that surprises nobody” and are always resolved in favor of the fossil data. Skobelin, Sharapov, and Bugayov, “Deliberations of state and ways of Perestroika in geology,” in Barto-Kyriadkidis, Critical Aspects of Plate Tectonics Theory (Athens: Theophrastas Publications, 1990), p. 25 (italics added), as quoted in Woodmorappe, The Mythology of Modern Dating Methods, p. 27.
19. Lewin, at p. 235. (“When Harris and White first examined the pig fossils, in 1974, their initial impression was that a 2.61 million-year date was correct for the KBS Tuff.”); Lubenow, at p. 262. 20. Lubenow, at p. 265. The fission-track dating of the KBS tuff at first confirmed Fitch and Miller, then later, remarkably, confirmed the younger date. Every time an atom of radioactive U-238 atom explodes, its nucleus is torn in half and the two halves are propelled through the crystal in opposite directions at great speed, boring a tiny tunnel through the crystal lattice. The idea behind fission track dating is to count the number of these fission tracks in the zircon crystal. When researchers Tony Hurford and Andy Gleadow did this on KBS tuff crystals, they independently reached a result of 2.62 m.y. plus or minus .40, a stunning confirmation of Fitch and Millers number of 2.61 m.y. plus or minus .26. (Lewin, at p. 231.) Later, after the tide of opinion had turned against Fitch and Miller, Hurford and Gleadow re-counted the fission tracks and came up with 1.8 m.y., a stunning confirmation of Garniss Curtis’ results. After they changed their tune, they claimed that Fitch had subtly pressured them to reach a number that confirmed his radiometric results, but it is hard to avoid the conclusion that, later on, the subtle pressure to support Curtis and McDougall became just as great. (Lewin, at 243-247.) Hurford stated, “[Y]ou can bias your results 10 percent either way, easily. You go crystal by crystal, and you begin to see where the rolling average is going. If you need the count to be higher with the crystal you’re working on, so that it will fit in, you might include something that is a doubtful track. If you want the count to be lower, you don’t include it.” (Lewin, at p. 246.) But, of course, Hurford and Gleadow changed their results not by ten percent, but by more than 30%.
21. Lewin, p. 235.
22. As a sadder and wiser Richard Leakey pointed out, “One realizes that even in the most pure of sciences, which geophysics should be, there is a potential to identify careers, status, and results—and there’s a strong political element, too.” Lewin, at p. 251.
23. The problem of “scatter” has not been ameliorated by more sophisticated equipment. “Improved laboratory techniques and improved constants have not reduced the scatter in recent years.” Waterhouse, J. B. “Chronostratigraphy for the World Permian,” Contributions to the Geologic Time Scale (American Association of Petroleum Geologists Studies in Geology no. 6, 1978), p. 316, as quoted by Woodmorappe, Studies in Flood Geology, p. 147.