In 2002, I published a report in the journal Radiocarbon. The report presented a hypothesis that pertains to the dates of some events in the Ancient Near East. Afterwards, a response to the report was published by Sturt W. Manning and co-authors. Following are some remarks on that response.
Synopsis of the hypothesis
The hypothesis concerns radiocarbon dating. The central idea is that most radiocarbon dates from the Ancient Near East are substantially too early. The cause of the hypothesized radiocarbon discrepancy is degassing by the Mediterranean Sea.
After the end of the last ice age, much of the Mediterranean Sea is known to have stagnated. During the stagnation, much of the 14C in the stagnating waters decayed, due to normal radioactivity. Additionally, the carbon content of the sea was greatly increased, due to inputs of organic matter and calcium carbonate (which contains no 14C). None of this is disputed by researchers in the field.
The Mediterranean Sea comprises two main basins: the Eastern Mediterranean and the Western Mediterranean, both of which are a few kilometers deep (see Figure 1 of my report). The two basins are separated by the Sicilian–Tunisian Strait, which is only 350 m deep. As the stagnation ended, the stagnant waters in the Eastern Mediterranean must have been upwelled, in order to pass through the Strait—else they would still be stagnant. As those waters upwelled (not necessarily to the surface), they depressurized, which would have led to the release of some of their 14C-deficient carbon.
The central question concerns when the stagnation ended. The most-commonly cited date for the end is about 4500 BC. My report argued that although the ending might have begun in 4500 BC, the ending was gradual, and that it took until about the year 0 for all the stagnated waters to be fully flushed out and the modern circulation to commence.
My report further suggested that during the time stagnant waters were being upwelled (4500–0 BC), some of the 14C-deficient carbon was released at the surface and that this might have affected the radiocarbon ages for organisms downwind. The report presented little evidence that such occurred—this was presented only as a hypothesis.
Response of Manning et al.
A response to my report was published by Manning et al. (reference below). The Response correctly noted an error in one sentence of my report. I posted a corrigendum about this on my web site.
The Response also claimed to find many other errors in my report. What follows critiques a representative selection of those claimed errors. The Response is divided into four main sections; that division is followed here.
Manning et al.: Section 1
Section 1 begins as follows: “Keenan (2002:225) claims that 14C ages are too old for the Mediterranean region from the "earliest historical times until the mid-second millennium BC" (sic—the earlier Holocene is meant)”. The Holocene is the period since the end of the last ice age (roughly, comprising the past 10000 years). Clearly, my report did not mean what the Response says: the Mediterranean was stagnant during the early Holocene, and so could not have been degassing then. Historical times began with the invention of writing (by definition)—which occurred late in the fourth millennium BC—and it is early historical times that my report addresses. Thus Manning et al. here seem to be confused about the central mechanism of the hypothesis.
My report presented a list of references for radiocarbon dates from the Ancient Near East that are a century or more earlier than the dates advocated by archaeo-historians. The Response claims to discredit this list, but it actually treats only one of the references: Bonani et al. (2001).
Bonani et al. presented radiocarbon dates from the pyramids. The Response argues that those radiocarbon dates are consistent with the dates advocated by archaeo-historians. The argument, however, is incorrect: there is an error in the statistical analysis, which would be obvious to anyone who knew statistics at the level taught in introductory courses. It is difficult to believe that none of the seven authors of the Response noticed the error. When the analysis is done correctly (see below), it shows that the radiocarbon dates are much earlier than the dates given by archaeo-history.
On 9 October 2003, I sent an e-mail to S.W. Manning and one of the co-authors, P.I. Kuniholm, pointing out the error in their analysis of the pyramid dates. Below is the relevant portion of that. (To relate this to radiocarbon dates, note that the probability that a normal measurement is more than one standard deviation greater than the average is about one sixth; i.e. the probability is about the same as having an ordinary six-sided die come up 6.)
To see the error in that analysis, imagine rolling an ordinary six-sided die. Suppose that the die comes up 6; this is reasonable--i.e. nothing special. Now roll the die eleven times. Suppose that the die comes up 6 ten out of eleven times. Obviously something is wrong (e.g. the die is loaded). According to the analysis of Manning et al., however, because each of the individual rolls of the die is reasonable on its own, then having the die come up 6 ten out of eleven times is also reasonable. You don't need any training in statistics to see that this is erroneous.
Manning did not respond. Yet in 2006, Manning published a paper claiming that the pyramid radiocarbon dates “provide no evidence at all for claims of hypothetical 100–300 years too early offsets in Mediterranean radiocarbon ages”.
The Response also discusses the archaeological site of Amarna, in Egypt. The discussion demonstrates that 14C dates for the site are consistent with archaeo-historical dates. The archaeo-historical dates, though, are around 1340 BC. Here is what my report says, in the first paragraph: “Too-early 14C dates occur from the earliest historical times until the mid second millennium BC”. Thus the Amarna dates are not evidence against the hypothesis. Additionally, Amarna is about 400 km inland from the Mediterranean (across desert terrain); so even if there were some weak marine degassing as late 1340 BC, its effect on the site would be attenuated, perhaps to nothing.
Also in 2006, Manning and six co-authors (including B. Kromer, one of the co-authors of the 2002 Response) published a report about the date of the (Minoan) eruption of the volcano Thera, in Greece. I submitted a comment to the publishing journal, pointing out several errors in their report. Below is a relevant portion of that comment. (The statistical analysis herein is a more precise formulation of the dice counter-argument; note that 0.159 is the precise probability that a normal measurement is more than one standard deviation greater than the average.)
… radiocarbon dates taken in and downwind of the Mediterranean are often substantially earlier than corresponding archaeo-historical dates … Manning et al. (2002) claimed to argue against this, but they ignored all except one reference, and their rebuttal of that reference is statistically incorrect: e.g. the probability that ≥k of n 14C dates are >1σ too early is about Σi≥k C(n,i)·0.159i·0.841n-i.
… there is also more recent work [showing too-early radiocarbon ages: see Wiener (2006) and Anderson (2006)].
The authors replied as follows.
[Keenan] repeats an assertion he made several years ago alleging that Mediterranean 14C ages in the BC era (he now qualifies this, conveniently, for no stated reason, as until the ‘mid second millennium B.C.’) are too old…. This claim has already been shown to have no merit, and all available evidence in fact contradicts it [Manning et al., 2002].
The reply is plainly untruthful and misleading. The journal, however, accepted it, and similar replies to some of the other errors that my comment described—without doing peer review; my comment was then rejected.
It is worth considering the example of the pyramid dates using the statistical formula in my submitted comment. The Response presented 25 radiocarbon dates, 20 of which are at least half a standard deviation earlier than the archaeo-historical date. If radiocarbon and archaeo-history were in agreement, then the probability of that occurring by chance is < 0.000005. This bound is extremely conservative; the real probability is much lower. Also, the bound holds even if a few of the earliest radiocarbon dates are discarded as outliers.
It should be noted that the foregoing would be obvious to anyone familiar with radiocarbon, and that three of the authors of the Response (Kromer, Levin, Reimer) have been doing research in radiocarbon for decades.
The Response further states “Keenan has carefully avoided citing any of the other studies that have found that, in general, Mediterranean region 14C dates usually agree perfectly well with the relatively secure early historic dates (e.g. Bruins et al. 2003; Hassan and Robinson 1987; Weninger 1990, 1997; Betancourt and Lawn 1984)”. I have written a separate exposition about that statement.
Manning et al.: Section 2
The Response claims the “Mediterranean stagnation ended by ∼5000 14C yr BP with increased overflow to the Atlantic as observed in the sedimentology and in the planktonic δ13C of a series of cores east and west of the Gibraltar sill (Vergnaud-Grazzini et al. 1989), not the 1000–0 BC quoted from this same study by Keenan”. Here is what Vergnaud-Grazzini et al. say (p.314): “[this study suggests] a varying intensity of the Mediterranean outflow, with slower episodes between 18 000 to 16 000 years [ago], 14 000 to 11 000 years [ago], 9000 to 3000 years [ago]”. The accompanying figures indicate that the end date for the slow episode during 9000–3000 years ago is a rough estimate and the episode might have actually lasted until 2000 years ago. In other words, the statement in my report about “roughly 1000–0 BC” is fine, and the claim of the Response is contradicted by the cited work.
As discussed above, the crucial issue for the hypothesis is determining when the stagnation was completely ended. Section 2.3 of my report is devoted to discussing that end date. The Response ignores all the evidence presented in the section, except for the evidence of Vergnaud-Grazzini et al.
(Note that although the stagnation might have fully ended during 1000–0 BC, significant degassing of very 14C-deficient carbon would be expected to have ended somewhat earlier, because the ending was gradual. This is discussed in my report.)
The Response also cites research of Siani et al. (2001) to claim that the sea-surface carbon “returned to near modern values by 6000 14C [years before present]”. S.W. Manning and I had an e-mail discussion about this before the Response was accepted for publication, in February 2003. Here is a quotation from one of my e-mails in that discussion.
Siani et al. present data from a single core taken from the southwestern Adriatic. It is well known that western Adriatic surface water is not Mediterranean surface water (see Figure 1 of my [report], which shows the area as hatched for that reason)….
(As shown in Figure 1, at the site that Siani et al. studied, the surface water is substantially outflow from a river—the River Po; it is not regular Mediterranean surface water. Siani et al. effectively acknowledge this [p.1918c], but do not consider the issue further.)
Manning replied to me, but did not attempt to rebut the point. Also in February 2003, I noted the point in e-mail to another author of the Response, P.J. Reimer. Thus both Manning and Reimer were aware, before the Response was accepted for publication, that their claim about the research of Siani et al. was at best implausible.
A copy of the e-mail discussion that Reimer and I had is here. In the discussion, Reimer demonstrates that she does not understand the hypothesis. Additionally, she cites results from other scientific papers as evidence against the hypothesis, but those papers do not say what she claims they do. Moreover, if they did say what she claims, then their results would be evidence for the hypothesis. I tried to explain some of the issues, but eventually gave up. The last sentence from my final e-mail shows my exasperation: "It is amazing that you … would publish a reply to my paper without first reading either my paper or the other papers that you cite against it".
The Response further cites studies indicating that a sea-surface deficiency in 14C might not force a deficiency in the surface air. Those studies, however, did not consider cases where 14C-deficient water was being upwelled. Yet upwelling is crucial for the hypothesis, because upwelling leads to depressurization of the dissolved carbon (which obviously could lead to increasing the amount of the carbon degassed into surface air). The authors' citing of non-upwelling studies thus implies either a failure to understand the basics of the hypothesis or an ignoring of such.
More can be said about this. Manning and I also discussed the degassing issue in February 2003. In the discussion, it became clear that Manning had not understood that upwelling was crucial for the hypothesis. In other words, Manning was the lead author of a (submitted) response to my report, while failing to understand the basics of the mechanism hypothesized in the report. In the discussion, I pointed out the cruciality of upwelling to Manning. Again, Manning replied and did not attempt to rebut the point. In other words, Manning knew that what the Response was saying here was invalid, prior to the Response being published.
The Response additionally says “In Levin et al. (1987) the difference in atmospheric Δ14C between the Northern Hemisphere and Neumayer Station on the Antarctic coast of the Weddell Sea was reported as −11‰ (equivalent to 88 yr, not 175 as claimed by Keenan)”. In fact Levin et al. (section 4.2) report that they observed a “Δ14C depletion at 70°S of about −20 to −25‰”. Since −21‰ is equivalent to 175 years, my report's claim is obviously valid. Note that Levin and her first co-author, Kromer, are both co-authors of the Response.
One study cited in my report is that of Rozanski et al. (1995). The study was of done in Ecuador, downwind from where upwelling in the South Pacific has occurred (the upwelling is related to El Niño). Rozanski et al. reported “direct evidence for regional reduction of 14CO2 levels due to gas exchange with 14C-[deficient] … surface ocean in the upwelling region”. This is the mechanism that is invoked by the hypothesis; modern direct observation of such degassing provides important support for the hypothesis.
The Response states “We disagree with Keenan's interpretation of the Rozanski et al. (1995) data as showing an atmospheric response to outgassing of old Pacific waters during an El Niño event”. The interpretation, however, is not mine. It is that of Rozanski et al.
Rozanski's first co-author, Levin, is a co-author of the Response. Thus Levin is disagreeing with her own, previously-published, interpretation. What justification does the Response give for such disagreeing? The justification is a study at another site, which did not show the same effect as the Ecuadorian site used by Rozanski et al. The other site, however, is in Venezuela. It is not downwind from the South Pacific; rather, it is downwind from the North Atlantic. Thus, it has no bearing on the study of Rozanski et al.
Note too that even if the other site were in a relevant location, that would not have entirely negated that study of Rozanski et al. Rather, that would have meant there was a discrepancy between the two sites, which required investigation.
Manning et al.: Section 3
My report contained a short Excursus, which strongly criticized some tree-ring studies done in Anatolia (roughly, modern Turkey). Section 3 of the Response is essentially a rebuttal to that Excursus.
Before publication of my report, I sent a copy of the Excursus to the principal investigator of the tree-ring studies, P.I. Kuniholm, asking for his comments. In Kuniholm's reply, one of his comments concerns what are known as “D-scores”. The D-score has been used frequently by Kuniholm and co-workers as a statistical test. D-scores are not used in statistics, though, and have no mathematical derivation; they have been used only in some tree-ring studies. The Excursus noted that D-scores are “meaningless”. In his reply, however, Kuniholm claimed that he had shown D-scores to one of the world's leading statisticians, Sir David Cox, who had “pronounced himself perfectly happy” with D-scores.
In fact, not only are D-scores meaningless, but this would be obvious to anyone with even a cursory understanding of statistics. In other words, Kuniholm's reply is dishonest.
Prior to receiving Kuniholm's reply, I had concluded that Kuniholm was incompetent (based upon reading some of his publications). It had not occurred to me that Kuniholm was dishonest. After receiving Kuniholm's reply, though, I decided to thoroughly review Kuniholm's work on Anatolian tree-rings. The primary conclusion of the review that I then wrote is that the work “has invalidating flaws, which are obvious upon inspection”. The review includes a point-by-point rejoinder to Section 3 of the Response, and also a detailed discussion of D-scores. (See further below.)
After I wrote the review, one of Kuniholm's Ph.D. students, M.W. Newton, finished and published her dissertation. The dissertation acknowledges that D-scores are “not subject to the rules of statistical probabilities”, and thus essentially accepts my report's criticism. Note that Newton is a co-author of the Response. In 2006, Kuniholm retired; in 2007, Manning, Newton, and co-workers announced that they no longer used D-scores.
Manning et al.: Section 4
Section 4 discusses radiocarbon measurements from Anatolia. The discussion demonstrates that there could not have been a radiocarbon deficiency in central Anatolia during most relevant ancient times. The Response says that the Anatolian measurements were “inexplicably ignored” by my report. The hypothesis presented in my report, though, is about the region in and downwind of the Mediterranean. This is stated in the title; this is stated throughout the report; this is obvious from the proposed mechanism. Anatolia is not downwind from the Mediterranean (it is downwind from the Black Sea). Hence radiocarbon dates from Anatolia would not have been affected by the proposed mechanism.
My report contains a figure showing wind patterns over the Mediterranean, which illustrates the above. The figure is for wintertime—the main growing season of Egypt—and the references for the figure demonstrate that summertime winds over Anatolia are similar. Even without the figure and references in my report though, the authors of the Response would surely be aware that Anatolia is not downwind from the Mediterranean, because some of them have worked in Anatolia for many years and also done climatic studies of the region. The Response has this to say on the issue of being downwind: “[Anatolia is] surely as Mediterranean as anywhere else [Keenan] lists!”. I.e. a purely rhetorical treatment.
The primary conclusion from the above is that the Response is obviously a fraud. For me, this was confusing—how could the authors expect to get away with a fraud that is so obvious? I duly filed a formal allegation of research misconduct against Manning, who was then at the University of Reading; I also filed an allegation against Kuniholm, who was then at Cornell University.
The fraud allegations that I filed were specifically for the tree-ring work of Kuniholm and Manning. As noted above, I had written a review of Kuniholm's work on Anatolian tree rings: that work constitutes a large majority of Kuniholm's career. Although the trees for the work were sourced in Anatolia, the laboratory where most of the research was done is at Cornell. Kuniholm was the head of the laboratory until retiring in 2006; he was then replaced by Manning. Prior to 2006, Manning had been co-working with Kuniholm.
The University of Reading refused to investigate my allegation. I was told that the university had no procedures for investigating such allegations, because their professors always acted with integrity.
Cornell University conducted a preliminary inquiry. Kuniholm was asked for the names of colleagues who would examine my allegation. Kuniholm then named four such colleagues—two of whom had done work with him on tree rings in Anatolia. Each of the colleagues said that Kuniholm was innocent. The report from Cornell's inquiry hence concluded that misconduct had not occurred.
After that, I understood what the Response was really about. The authors were, in effect, sending me a message: they can get away with publishing virtually anything, and there is nothing that I can do about it. Several people have asked me why I have not published a rejoinder to the Response. The answer should be clear. I assume that the authors would just respond with additional fraud.
Finally, I am not the only person to have accused Manning of fraud. In 2004, one of the world's leading Egyptologists, Manfred Bietak, published a review of a book that Manning wrote (references below). The review is over 12 000 words long, and details repeated instances of dishonesty and misrepresentation in the book. It is notable that Bietak and I came to our conclusions independently, and only found out about each other's review afterwards.
Since the publication of the Response, some of the authors have published more fraudulent work. The 2006 article by Manning and the 2006 article by Manning, Kromer, and five others were discussed above. And Reimer was the first co-author of an article that analyzed the effects of upwelling 14C-deficient water off the coast of Brazil but ignored degassing (Angulo et al., 2007).
Although my review of the Anatolian tree-ring work was not formally published, it was available on my web site, and it was widely viewed. Some tree-ring researchers told me privately that they were shocked by reading about what Kuniholm et al. had done. In 2009, Manning co-authored a “reappraisal” of one of the major findings in Anatolian tree-ring work (dating an ancient gateway). The reappraisal concluded that the original finding was incorrect and acknowledged that my criticism of that finding was valid (Griggs & Manning, 2009).
In 2010, Manning, Kromer, and five others published a paper in which they admitted that there is a “large difference between 14C-based dates and chronologies for the mid-2nd millennium BC Aegean world (e.g. Manning 1999; Kutschera and Stadler 2000; Bronk Ramsey et al. 2004; Friedrich et al. 2006; Manning et al. 2006, 2009a; Wild et al. 2010), versus the much later calendar date estimates made by some archaeologists and art historians based on the traditional/conventional interpretation of comparisons of artifact or stylistic linkages between the Aegean, the east Mediterranean, Egypt, and the Near East (e.g. Warren and Hankey 1989; Bietak 2003; Bietak and Höflmayer 2007; Wiener 2003, 2006; Bruins 2010 …)”. This admission contrasts with the claim of Manning et al. (2002) that “14C dates usually agree perfectly well with the relatively secure early historic dates” (discussed above in Section 1).
Manning et al. (2010) also present further measurements from central Anatolia, and then claim that they “observe no evidence compatible with claims (Keenan 2002) of a purported depletion of 14C levels in the Mediterranean”. Similarly, Kromer et al. (2010) claim that the measurements in central Anatolia “leave no room for a purported (Keenan 2002) regional depletion of 14C ages in the eastern Mediterranean”. In both cases, the authors know that central Anatolia is not downwind from the Mediterranean, and thus their measurements are irrelevant for the degassing hypothesis (as discussed above in Section 4).
Evidence for the hypothesis
Good evidence for the hypothesis comes from a deep-sea sediment core recovered 135 km southeast of Thera. The core contains volcanic ash from Thera's eruption: the ash was mixed with surface-plankton remains whose radiocarbon age is at least 400 14C years older than the age of the eruption (Mercone et al., 2000). (Mercone et al. attribute the discrepancy to enhanced bioturbation that was induced by additional carbon in the ash; the ash, however, contains no carbon, and ashes from several other volcanoes in the Mediterranean have not enhanced bioturbation [Siani et al., 2001].) This implies that, at the time of the eruption, the sea-surface carbon was seriously deficient in 14C. There does not seem to be a plausible scenario that could account for that other than upwelling of 14C-deficient water.
❧ On 2009-10-03, a link to this web page was sent to the authors of the Response: Sturt W. Manning, Mike Barbetti, Bernd Kromer, Peter Ian Kuniholm, Ingeborg Levin, Maryanne W. Newton, Paula J. Reimer. To date, I have received no reply.
Angulo R.J., Reimer P.J., de Souza M.C., Scheel-Ybert R., Tenório M.C., Disaró S.T., Gaspar M.D. (2007), “A tentative determination of upwelling influence on the paleo-surficial marine water reservoir effect in southeastern Brazil”, Radiocarbon, 49: 1255–1259.
Betancourt P.P., Lawn B. (1984), “The Cyclades and radiocarbon chronology”, The Prehistoric Cyclades (editors—MacGillivray J.A., Barber R.L.N.) 277–295 (Department of Classical Archaeology, University of Edinburgh).
Bonani G., Haas H., Hawass Z., Lehner M., Nakhla S., Nolan J., Wenke R., Wölfli W. (2001), “Radiocarbon dates of Old and Middle Kingdom monuments in Egypt”, Radiocarbon, 43: 1297–1320.
Friedrich W.L., Kromer B., Friedrich M., Heinemeier J., Pfeiffer T., Talamo S. (2006), “Santorini eruption radiocarbon dated to 1627–1600 B.C.”, Science, 312: 548. doi: 10.1126/science.1125087. [This is a companion to Manning et al. (2006), with Kromer as the first co-author.]
Griggs C.B., Manning S.W. (2009), “A reappraisal of the dendrochronology and dating of Tille Höyük (1993)”, Radiocarbon, 51: 711–720.
Hassan F.A., Robinson S.W. (1987), “High-precision radiocarbon chronometry of ancient Egypt, and comparisons with Nubia, Palestine and Mesopotamia”, Antiquity, 61: 119–135.
Keenan D.J. (2002), “Why early-historical radiocarbon dates downwind from the Mediterranean are too early”, Radiocarbon, 44: 225–237.
Keenan D.J. (2006), “Anatolian tree-ring studies are untrustworty”, (manuscript). [This is an updated version of the review that documented my allegation of research misconduct.]
Kromer B., Mannin, S.W., Friedrich M., Talamo S., Trano N. (2010), “14C calibration in the 2nd and 1st millennia BC—Eastern Mediterranean Radiocarbon Comparison Project (EMRCP)”, Radiocarbon, 52: 875–886.
Levin I., Kromer B., Wagenbach D., Münnich K.O. (1987), “Carbon isotope measurements of atmospheric CO2 at a coastal station in Antarctica”, Tellus, 39B: 89–95.
Manning S.W. (1999), A Test of Time (Oxford: Oxbow).
Manning S.W., Barbetti M., Kromer B., Kuniholm P.I., Levin I., Newton M.W., Reimer P.J. (2002), “No systematic bias to Mediterranean 14C ages”, Radiocarbon, 44: 739–754. [This was published in July 2003, notwithstanding the nominal date.]
Manning S.W., Bronk Ramsey C., Kutschera W., Higham T., Kromer B., Steier P., Wild E.W. (2006), “Chronology for the Aegean Late Bronze Age 1700–1400 B.C.”, Science, 312: 565–569. doi: 10.1126/science.1125682.
Manning S.W., Griggs C., Newton M., Wazny T. (2007), Summary and Discussion of Procedures at The Malcolm and Carolyn Wiener Laboratory for Aegean and Near Eastern Dendrochronology. [This states that the authors no longer use D-scores (p.3).]
Manning S.W., Kromer B., Bronk Ramsey C., Pearson C.L., Talamo S., Trano N., Watkins J.D. (2010), “14C Record and wiggle-match placement for the Anatolian (Gordion area) juniper tree-ring chronology ~1729 to 751 cal BC, and typical Aegean/Anatolian (growing season related) regional 14C offset assessment”, Radiocarbon, 52: 1571–1597.
Mercone D., Thomson J., Croudace I.W., Siani G., Paterne M., Troelstra S. (2000), “Duration of S1, the most recent sapropel in the eastern Mediterranean Sea”, Paleoceanography, 15: 336–347. [Section 3.2.4 describes the early radiocarbon date for the eruption of Thera.]
Rozanski K., Levin I., Stock J., Guevara Falcon R.E., Rubio F. (1995), “Atmospheric 14CO2 variations in the equatorial region”, Radiocarbon, 37: 509–515.
Siani G., Paterne M., Michel E., Sulpizio R., Sbrana A., Arnold M., Haddad G. (2001), “Mediterranean Sea surface radiocarbon reservoir age changes since the last glacial maximum”, Science, 294: 1917–1920. doi: 10.1126/science.1063649.
Vergnaud-Grazzini C., Caralp M., Faugères J.-C., Gonthier E., Grousset F., Pujol C., Saliège J.-F. (1989), “Mediterranean outflow through the Strait of Gibraltar since 18 000 years BP”, Oceanologica Acta, 12: 305–324.
Weninger B. (1990),“Theoretical radiocarbon discrepancies”, Thera and the Aegean World III (editor—Hardy D.A.) 3: 216–231 (London: Thera Foundation).
Weninger B. (1997), Studien zur Dendrochronologischen Kalibration von Archäologischen 14C-Daten (Bonn: Dr. Rudolf Habelt).