Looks like you are using an old version of Internet Explorer - Please update your browser
Originally published in Creation 11, no 2 (March 1989): 30-34.
In 1979, scientists Eddy and Boornazian cautiously announced that their studies of solar measurement records from Greenwich Observatory in England, and the US Naval Observatory in Washington, conclusively showed that the sun was shrinking. Its diameter was decreasing at a rate of almost six feet per hour.1
There were potentially astounding implications. The announcement by Eddy and Boornazian with respect to the age of the sun (and hence the solar system), along with the apparent conflict with previously held ideas about how the sun produces its heat and light, did not go unnoticed. A vigorous healthy debate among solar astronomers began.
After their own analyses of the Greenwich and Washington data, plus comparisons of solar eclipse records, and consideration of photographic and other relevant evidence, many colleagues agreed with Eddy and Boornazian but the consensus seemed to be that the shrinkage rate was less than half that initially suggested.
However, not all scientists agreed that the evidence indicated the sun is shrinking. A number disputed the reported figures and presented results of their own that seemed to indicate no shrinkage. Among them was Irwin Shapiro of the Massachusetts Institute of Technology. He reported a series of figures for the sun’s diameter that he had calculated from studies of transits of the planet Mercury.
For more than two centuries astronomers had been studying the exact orbit of Mercury. Occasionally Mercury passes directly between the earth and the sun, so that it crosses our direct line of sight to the sun. Viewed from earth, Mercury appears to cross the face of the sun from one side to the other, and this is called a transit.
Irwin Shapiro had collected a whole series of records of observations of the transit of Mercury. He realized that if he put into his computer the time at which the transit started, and the time at which the transit finished, he could use these records as a series of measurements of the sun’s diameter. If the sun’s diameter had been larger in the past, then the transit time of Mercury should have been longer then than it is today.
So he analysed his collection of data. He concluded that the records from 23 transits of Mercury between 1736 and 1973 indicated that there had not been any statistically significant change in the sun’s diameter over those 237 years. He reported his findings in the journal Science in April 1980.2
But a closer look at Shapiro’s results, and when statistical error margins were applied to each of his data points, then it was clear that Shapiro’s analysis of the transit of Mercury data could not definitely rule out the possibility that there was some shrinkage of the sun. All Shapiro could really say was that he couldn’t detect any shrinkage if shrinkage was indeed occurring. Yet while his results showed no indication of any significant change in the diameter of the sun, his regression analysis yielded a decrease in solar diameter of under 0.2 second of arc per century at a confidence limit of >90%. Thus it could easily be argued that Shapiro’s results are still comparable with Dunham et al.’s approximate 0.2 second of arc per century shrinkage rate based on records of the 1715 and 1979 solar eclipses,3 and Howard’s 0.5 second of arc per century shrinkage rate from 50 years (1930–1980) of solar photography.4
But three English scientists, Parkinson, Morrison and Stephenson also looked at the Greenwich data and their own set of Mercury transit data as well as the solar eclipse data, and they too came to the conclusion that there was no detectable shrinkage of the sun during the past 250 years.5 However, they suggested that there is some evidence for the occurrence of periodic changes in the sun’s diameter of about 0.02% on a time-scale of 80 years. They also felt that Eddy and Boornazian’s findings were the result of instrumental and observational defects (they outlined these in detail) rather than real changes.
Next to join the fray was German scientist Wittmann, who pointed out that the distinguished eighteenth century astronomer, Tobias Mayer, had made a series of highly reliable observations of the sun between 1756 and 1760. Wittmann proceeded to analyse Mayer’s 129 transit observations and declared that there was excellent agreement with more recent photoelectric transit observations, thus lending no support whatsoever to claims that the sun is shrinking.6
Subsequently, in a valiant effort to decisively resolve this debate over whether the sun is shrinking or not, Ronald Gilliland undertook a comprehensive reassessment of all the five data sets of apparent solar diameter measurements published by the authors participating in the debate. His results end conclusions were published in The Astrophysical Journal of the American Astronomical Society in September 1981.7 In his summary, Gilliland reported that his analysis of the five different data sets, including the meridian circle (Greenwhich and Washington) observations, the timings of transits of Mercury, and the durations of total solar eclipses, consistently suggested the presence of a 76-year modulation (cycle of variation) of the solar radius. He further suggested that the last solar radius maximum occurred in 1911, and that the half-amplitude of variation (half the change in radius between the minimum and maximum of the cycle) is approximately 0.2 second of arc per century or 0.02% of the solar radius.
Interestingly, Gilliland was also mildly critical of some of his colleagues’ handling of the data. For example, he pointed out that in his analysis he had not thrown out portions of the Greenwich data as suggested by Parkinson, Morrison and Stephenson to account for possible systematic errors resulting from instrumental and observational changes. He reasoned that to remove certain sections of the data set which showed discontinuities correlated with instrumental changes tended to introduce further biases into the data sat. He concluded that the reader should be warned of uncertainties which exist in the individual data sets as Parkinson, Morrison, and Stephenson had done, but subjective removal of certain sections to support the Parkinson et al. premise that the solar diameter has been constant over the past 250 years should also be viewed with caution.
In deriving his results, Gilliland commented that the remarkable agreement between independent data sets and combinations of sets in predicting the maximum twentieth century radius near l910 with a cycle time of about 76 years is the strongest evidence supporting a cyclic solar radius change. Furthermore, the most noticeable feature common to all five data sets is a decrease in solar radius from l910 to the 1940s. He noted that without exception the analyses show this twentieth century feature. The fact that this cyclicity is clearly evident in both the Greenwich and Washington observations lends support to his contention that these series of measurements still have some validity despite Parkinson, Morrison and Stephenson’s attempt to downplay their significance due to claimed instrumental and observational defects.
Gilliland was also bold enough to admit that since stellar evolution theory predicts that the sun should increase in size with increasing age (i.e. the sun’s diameter should be increasing), any decrease is quite significant. To be sure, he said, the discrepancies between independent data sets—for example, a clear long-term decreasing trend in the Greenwich measurements reported by Eddy and Boornazian in 1979 and the lack of a trend in the Mercury transit data of Shapiro (1980)—makes simple interpretations problematic. But Gilliland maintained that in the partially justified, but perhaps overzealous criticism of the early Eddy and Boornazian claims there is the distinct possibility that much smaller but still fundamentally important long-term trends were being inadvertently disclaimed. He then noted, as we have already done above, that the equatorial trend derived from Mercury transits by Parkinson, Morrison and Stephenson over the interval 1723–1973 precisely agrees with the polar radius decrease of almost 0.2 second of arc per century over the interval 1715–1979 derived from observations of total solar eclipse path widths by Dunham et al.
Even more telling is the fact that even though Parkinson, Morrison and Stephenson argued that the horizontal Greenwich measurements were not reliable before 1854 or after 1915 because of instrumental and observational inadequacies, analyzing only the horizontal Greenwich data from 1854–1914 yields a long-term decrease trend of just over 0.3 second of arc per century. Thus Gilliland claimed that the objective result from the Parkinson, Morrison and Stephenson (1980) paper should have been that the Mercury transit data support a long-term radius decrease of over 0.1 second of arc per century and that the most reliable portions of the Greenwich observations support a somewhat steeper decrease.
To quote Gilliland:
Thus we can conclude that a thorough analysis of all the available evidence clearly suggests a steady long-term decrease of the solar diameter (i.e. the sun is shrinking) at a rate of almost 0.2 second of arc (150 kilometers or 93 miles) per century or approximately 30 centimeters (less than one foot) per hour, superimposed upon a 76–80 year cycle of systematic increase and decrease over a range of 0.8 second of arc (600 km or 373 miles).
But Gilliland’s thorough analysis of the data and definitive conclusion have far from settled the debate. Not unexpectedly, both Stephenson and Parkinson responded to Gilliland’s reanalysis of the available data. Stephenson reported in Scientific American8 that his reanalysis (with the help of his colleagues) of the thirty transits of Mercury and six total solar eclipses between 1715 and 1979 had indicated only a negligible change in the sun’s diameter, calculated as a decrease of 0.16 ± 0.14 second of arc per century. This, Stephenson again suggested, was essentially a null result—the sun was not shrinking on a long-term basis, its diameter merely oscillating at regular intervals of about 80 years.
But this is not ‘essentially a null result’. More precisely worded, Stephenson’s conclusion should have been that the sun’s diameter is decreasing at a rate somewhere between 0.02 and 0.30 seconds of arc per century, a rate not incompatible with Gilliland’s suggested almost 0.2 second of arc per century. Even a rate of 0.02 second of arc per century amounts to a shrinkage of 15 km (more than nine miles) per century—hardly a near null result! Such reporting is not completely honest, and only confusing at best.
Only a year later (1983), Stephenson’s colleague, Parkinson, reported on his measurements made during the total solar eclipse of 1981 and claimed that together with a reanalysis of previous eclipse and Mercury transit measurements, the data confirmed that there is no evidence for any long-term change in the solar diameter, only increased support for an approximate 80 year cyclic variation.9 Because of the considerable accuracy in the reported observations and measurements of the 1929 eclipse, Parkinson concluded that therefore it appears that in 1924–25 the radius of the sun was approximately 0.5 second of arc larger than its average value, and equated this with a maximum in the approximate 80-year variation cycle that he and his colleagues had earlier deduced from the Mercury transit observations.
Independent confirmation of this conclusion came but two pages later in the same issue of Nature where Sofia et al. presented their fresh analysis of the numerous reports on the 1925 and 1979 solar eclipses.10 From different locations, observers had reported on the duration of the totality during the solar eclipses, and from this data Sofia and his colleagues found that the solar radius in 1925 was 0.5 second of arc or 375 kilometres (233 miles) larger than in 1979. And they also reported that although they had shown the solar radius decreased by 0.5 second of arc between 1925 and 1979, the sun’s size in 1925 was not significantly different from that in 1715. Thus they concluded that the solar radius changes are not a simple long-term uniform trend. They went on to claim that this is consistent with the null result (no shrinkage) from the transit of Mercury measurements which they claimed were less accurate than the eclipse results. Like Stephenson and Parkinson, they ignored the criticisms of Gilliland and his careful reanalysis of all the data sets, to claim that the Mercury transit data convincingly disproved the existence of a large long-term uniform decrease of the solar radius.
This of course raises the question as to which scientist or scientists one should believe? Similarly, whose mathematical analysis of the historical data should be believed? All the solar astronomers whose work we have cited have impeccable credentials. Their choice of mathematical manipulations of the observational data vary according to preference, yet they each promote their own conclusions with equal conviction. The layperson is left wandering in a maze and wondering how he can resolve the apparently conflicting evidence to arrive at the truth. While the debate continues to this day, a clearer picture is gradually emerging, one that still challenges the evolutionists’ multi-billion year age for the sun and solar system. This will be the subject of our third and final part of the series (published in Creation 11(3):40–43, June–August 1989).