Review Series to Celebrate Our 100th Volume
Motonori Matuyama and reversals of geomagnetic field
2024 年 100 巻 9 号 p. 491-499
詳細
2024 年 100 巻 9 号 p. 491-499
In 1929, Matuyama published his paper on the magnetization of mostly Quaternary volcanic rocks. In this paper, he described the results of paleomagnetic measurements of volcanic rocks from Japan and nearby areas and concluded that the latest transition of the magnetic field from reversed to normal state occurred in the early Quaternary. In the 1960s, two groups of scientists from the USA and Australia quite vigorously conducted studies of both magnetization and age of volcanic rocks. By about 1966, they completed the reversal timescale for the last 4 million years, which was to become the basis for many earth science studies. For easy reference, they suggested to call the most recent normal or reversed periods as Brunhes, Matuyama, Gauss, and Gilbert polarity epochs, with the names taken from the scientists who made very important contributions to paleomagnetism. Chron is now the official term for the epoch, and each chron is specified by a combination of a number and a character showing the polarity. However, the names of polarity epochs were already so popular that they are still quite frequently used in scientific papers. The Matuyama epoch is between 0.773 and 2.595 million years before present. Moreover, its lower limit is now used to define the start of the Quaternary.
The paper entitled “On the direction of magnetisation of basalt in Japan, Tyôsen and Manchuria” by Motonori Matuyama was published in the Proceedings of the Imperial Academy of Japan in 1929.1) The paper appeared to be one of those which did not receive much attention from scientific community at the time of its publication and became forgotten over time. However, it made an unexpected revival in the 1960s and became one of the most frequently cited papers from Japan for some time. We show what led to this dramatic reevaluation.
Matuyama’s paper reported the results of the measurements of remanent magnetization in volcanic rocks of Japan and nearby areas. First, Quaternary basalt rocks from Genbudô and Yakuno, both in Kinki district of Honsyû Island, were examined carefully. While the magnetization in the Yakuno rocks was directed to about 20°E and 50° down, which is similar to the present magnetic field direction (this state is called normal or N), the magnetization in Genbudô rocks showed south and upward direction, which is almost antipodal to the present field (reversed or R in the present terminology). The study was then expanded to include 36 volcanic rocks from other parts of Japan, Korea, and NE China. In Kyûsyû Island of Japan and the Tansen area of Korea, samples obtained from Quaternary rocks showed magnetic directions similar to the Yakuno site. On the other hand, in the rocks from Kissyû of Korea, which were somewhat older (but in the Pleistocene), the magnetic directions were similar to that in the Genbudô case. From these results, it was concluded that the magnetic field direction in this area changed from reversed to normal in the early Quaternary.
Measurements of older rocks showed somewhat strange results. Miocene volcanic rocks from these areas again formed two groups. However, the directions are quite different from the normal or reversed directions observed in Pleistocene rocks: one was with south and downward direction, while the other was with north and upward direction. Notwithstanding these differences, Matuyama concluded that there was a reversal of the geomagnetic field similar to the Quaternary case, as they formed almost an antipodal pair.
Figure 1 shows the results of his study, indicating the directions of remanences in each rock sample. How can we evaluate this paper in the development of paleomagnetism of the earth’s magnetic field and in the study of geomagnetic reversals in particular?
Magnetic directions of volcanic rocks observed by Matuyama.1) Vertical and horizontal directions correspond to inclination and declination of the natural remanent magnetization (NRM), respectively.
The fact that some rocks carry strong magnetization was known before measurement of such remanences actually began. A report by Alexander von Humboldt gives an example of the early recognition of magnetic effects due to volcanic rocks. He noted that a very large deflection was sometimes observed in the magnetic compass direction near volcanic rocks and attributed this effect to magnetization caused by lightning strikes.2) In later years, Humboldt became a great explorer and scientist and made significant contributions to various branches of science. His significant findings in geomagnetism include the discovery of geomagnetic storms and the measurements which showed that the intensity of the earth’s magnetic field systematically increased from the equator to higher latitudes.
Actual measurements of the remanent magnetization in rocks started in the early nineteenth century in Europe. Among these efforts, Macedonio Melloni measured a number of volcanic rocks and concluded that the magnetization in volcanic rocks was nearly parallel to the earth’s magnetic field at the site.3) In later studies of volcanic rock magnetization, Brunhes4) and Mercanton5) found that some of the rocks were magnetized in the direction almost opposite to the present magnetic field. They proposed that at the time the rocks were formed, the direction of the earth’s field was opposite to its direction today.
Matuyama’s paper appeared at such a time, but he was not the first person in Japan to measure the remanent magnetization of volcanic rocks. An earlier study by Nakamura and Kikuchi6) already confirmed Melloni’s result that volcanic rocks acquire magnetization in the earth’s magnetic field direction. What was new in Matuyama’s work was the careful examination of the ages of volcanic rocks with normal and reversed magnetizations. The age information available at that time was not very accurate, but he could argue that all reversely magnetized rocks were of early Pleistocene or older ages, which were older than the ages of normally magnetized rocks. From these observations, he concluded that the geomagnetic field changed from a reverse to a normal state in the early Pleistocene. In addition, he proposed that the existence of antipodal directions of magnetization in Miocene and older rocks indicate that polarity reversal may have also occurred in times older than Pleistocene.1)
Reports of the measurement of remanence in volcanic rocks became more and more abundant after Matuyama’s time; however, progress in the understanding of the geomagnetic reversals did not proceed in a straightforward way. Émile Thellier7) and Takesi Nagata8) conducted extensive experimental studies to clarify the fundamental properties of thermoremanent magnetization (TRM) in volcanic rocks. This was an essential step in understanding the natural remanent magnetization (NRM) of volcanic rocks. One of the remaining problems was the self-reversal of the TRM, which is a phenomenon where the direction of the TRM acquired is exactly opposite to that of the magnetic field in which the rock is cooled. Seiya Uyeda discovered such self-reversal in a volcanic rock collected at Mt. Haruna in Central Japan.9) Proving that self-reversal did not occur in that rock was necessary to convincingly conclude that a reversely magnetized rock was actually magnetized in a reversed geomagnetic field. In practice, this was quite difficult to show.
As self-reversal was a rare effect, it was hoped that polarity reversals may be shown to occur simultaneously at different places on the globe by collecting age data for many volcanic rocks with both normal and reversed remanences. However, even the occurrence of a reversal of the earth’s magnetic field in Pleistocene could not be validated so easily. The main difficulty was in determining the age of rocks with necessary accuracy (about 0.1 million years). Dating methods based on the decay of radioactive atoms were already available, but most of them were for the very young (carbon-14 method) or for the very old ages (rubidium–strontium or lead–lead methods). Potassium–argon (K–Ar) dating had a possibility to fill this gap. However, the presence of the atmospheric 40Ar plagued mass spectroscopy in the early 1950s, and accurate dating of rocks younger than about 10 million years was practically impossible.
John Reynolds of University of California at Berkeley discovered a big step in solving this problem by constructing an all-glass mass spectrometer. He showed that his system can be heated to high temperatures, and the contaminating atmospheric Ar can be degassed to a very low level.10) The possibility to use the K–Ar method for dating of the geomagnetic reversals became apparent to researchers at Berkeley. Among them, Brent Dalrymple joined the U.S. Geological Survey (Menlo Park) after getting a degree, where Allan Cox and Richard Doell (both from Berkeley) were already working on paleomagnetism of young volcanic rocks. Dalrymple built a dating facility in the USGS with a Reynolds-type mass spectrometer. Ian McDougall was a postdoc at Berkeley about the same time and did the same thing when he returned to his lab at Australian National University (ANU).
In 1963, the two groups started to report the paleomagnetic results of volcanic rocks with ages determined by the K–Ar method. The USGS team was the first to report their results,11) closely followed by the ANU team.12) At first, it appeared to these authors that N and R states alternate with about a 1-million-year duration. In one of their papers, the USGS team proposed that the most recent N and R periods of a roughly 1-million-year duration may be called as “polarity epochs” with the names of the scientists who made important contributions to the development of paleomagnetic studies. They suggested the names of Brunhes (N), Matuyama (R), Gauss (N), and Gilbert (R), for the most recent four epochs. Carl Friedrich Gauss was one of the greatest mathematicians and physicists who discovered the Gauss’s laws of the electric and magnetic fields. In his 1600 book, William Gilbert showed that the earth’s magnetic field is quite similar to the field produced by a uniformly magnetized sphere which he prepared for this study and concluded that the earth is a one giant magnet.13) However, further studies revealed that these epochs sometimes contain rather short periods of time in which the magnetic polarity was opposite to that of the epoch. Such periods were then given the name of “events,” such as Jaramillo (about 0.7 Ma) and Olduvai (1.9 Ma) events in the Matuyama reversed epoch.
In the following years, the USGS and ANU groups worked very hard to obtain paleomagnetic data from young volcanic rocks, backed by K–Ar age data. This appeared to be a big competition between the two teams that was taking place before the audience of the earth science community. Many of their papers appeared in Nature or Science, the two most prominent scientific journals. After a few years, their results seemed to converge as far as the epoch boundaries and major events are concerned. The reversal timescale for the last 4 million years or so was completed around 1966.14) Figure 2 demonstrates the rapid development of the reversal timescale. The influence of this establishment of geomagnetic polarity timescale was enormous. Reversals of magnetic directions were also found under the sea. Opdyke and others measured the magnetic remanences in deep-sea sediments collected by piston cores and used the reversal timescale obtained from land to calibrate the ages of each magnetic transition.15)
Early development of the geomagnetic polarity timescales. Black and white parts represent the times of normal polarity and those of reversed polarity, respectively. After Dalrymple.26)
The biggest application of the magnetic reversal chronology was the dating of the seafloor itself. In 1963, Fred Vine and Drummond Matthews of Cambridge University16) proposed that the seafloor is created at the ocean ridges, such as Mid-Atlantic Ridge, and spread out to left and right from the ridge. The rocks forming the seafloor erupt at the ridge with high temperature and cool to an ambient temperature. Through this process, the rocks acquire TRM; however, its direction is either normal or reversed depending on the polarity of the geomagnetic field at that time. The result of this process is the normally and reversely magnetized strips of seafloor forming a symmetric pattern against the ocean ridge. This is quite similar to the workings of a magnetic tape recorder.
This proposal was a combination of the seafloor spreading theory of Hess17) with the acquisition of magnetic remanence in the erupted volcanic rocks. The spreading produces normally and reversely magnetized strips of seafloor if the earth’s field reversed its polarity as shown by the studies of volcanic rocks on land. At first, the earth science community accepted their proposal with a lot of skepticism. However, using the magnetic data obtained by oceanographic surveys conducted by US institutions, Vine showed that the observed magnetic anomalies at the central part of the oceanic ridges displayed very convincing match with the anomalies obtained from the crustal magnetization model based on the reversal timescale of the last few million years.18) Researchers from Lamont Geological Observatory of Columbia University followed with the analysis of the oceanic magnetic field data collected from all the main oceans in the world and showed that Vine and Matthews theory can explain the magnetic anomalies observed in all these oceans.19) They could extend the reversal timescale to the entire Cenozoic using these analyses, assuming that the spreading rate is nearly constant at some of the ridges.
At about the same time, a lot of evidences were obtained from different branches of earth sciences, which can only be explained by large-scale motion of the earth’s surface. This not only caused the revival of continental drift theory of Alfred Wegener20) but also led to the establishment of the plate tectonics, which is now considered the most fundamental framework of the evolution and dynamic state of the earth’s structure.
The magnetic polarity timescale was so useful in dating various land and ocean basin geologic processes that different authors have revised and updated the timescale many times. Presently, the most popularly used compilation is that of Cande and Kent21),22) or their followers (e.g., Ogg, 202023)). In these timescales, chron replaces the name “epoch,” which starts from 1 and continues as 2, 3, etc., for the older ages, following the guidelines set by the International Subcommission on Stratigraphic Classification. Polarity events are also rephrased as subchrons. Each chron contains both normal and reversed periods of 1 million years or so. For instance, the former Brunhes epoch and the upper half of the Matuyama epoch are officially referred to as C1n and C1r. However, the use of epoch names was already so popular in the community that the names of Brunhes and Matuyama are almost always used in earth science papers.
The important properties of magnetic reversals are (1) that they are worldwide phenomena and (2) they are almost instantaneous. Actually, the transition from one polarity to the other takes an interval of a few thousand years, which is a very short time considering geological processes. Thus, magnetic reversals can be used as time markers. For instance, the Matuyama–Gauss polarity reversal defines the boundary between the Neogene and Quaternary. Moreover, recently the International Union of Geological Sciences adopted the Chibanian Stage in the middle of Quaternary, and the base of the Chibanian is set exactly at the Brunhes–Matuyama polarity reversal.24) Therefore, both the lower and upper limits of the Matuyama epoch serve as important time markers in the Cenozoic history.
In retrospect, Matuyama’s main contributions to the polarity timescale studies were that (1) he concluded that the most recent reversal occurred in the early Quaternary and (2) he suggested that reversals also occurred in earlier times such as the Miocene. The first point was quite appropriate, considering the fact that the Quaternary was still not quite well defined at the time. The second point needs some critical assessment, as it was based on some queer magnetic directions he observed in older rocks.
Later paleomagnetic studies established that average dipole direction coincides with the earth’s rotational axis and that such large deviations from the dipole directions occur only rarely (when the field is reversing its polarity or in some unusually large fluctuations called geomagnetic excursions). Such observations are also consistent with theoretical and numerical results of field generation models (dynamo theory), showing a very strong coupling of the dipole component with the earth’s rotational axis.25) Movement of the rotational axis is possible in a very long time interval, but such effect (polar wander) is quite small in the Cenozoic.
Matuyama’s queer magnetic directions (roughly northward and upward or southward and downward) in the rocks older than the Quaternary cannot be taken as characteristic of the magnetic field at that time. Their nearly antipodal directions may just be a coincidence without any significance. His statement, declaring that these directions probably indicate the existence of two opposed polarities in the Miocene, is perhaps premature. The amount and quality of data available at that time were not good enough to allow such conclusions.
In conclusion, Matuyama showed that geomagnetic reversal occurred in the early Quaternary based on a good argument. However, his suggestion that similar reversals occurred also in older ages, such as the Miocene, appears to be premature. We can take this representing his conviction that the reversals are not restricted to the Pleistocene but are common phenomena in the older ages. The later paleomagnetic and dating studies of the 1960s convincingly proved the fact that geomagnetic reversal occurred repeatedly in earlier geological time. Therefore, Matuyama’s conclusion as well as his speculation was true.
Edited by Yoshio FUKAO, M.J.A.
Correspondence should be addressed to: M. Kono, Yayoi-cho 2-4-4, Nakano-ku, Tokyo 164-0013, Japan (e-mail: masarukono@gmail.com).
This paper commemorates the 100th anniversary of this journal and introduces the following paper previously published in this journal. Matuyama, M. (1929) On the direction of magnetisation of basalt in Japan, Tyôsen and Manchuria. Proc. Imp. Acad. 5 (5), 203-205 (https://doi.org/10.2183/pjab1912.5.203).
[From Proc. Imp. Acad., Vol. 5 No. 5, pp. 203-205 (1929)]
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/100_pjab.100.031_3.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/100_pjab.100.031_4.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/100_pjab.100.031_5.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
