Spectrochimica

Acta,Vol.35B,pp.663 to670 0584-8547/80/1101-063$02.00/0

Ltd. 1980.F’rinted
@ PergamonPress in Great Britain

A history of atomic absorption spectroscopy

S. R. KOIRTYOHANN
Department of Chemistry and The Environmental Trace Substances Research Center,
University of Missouri, Columbia, MO 65211, U.S.A.

(Received 11 June 1980)

Abalraet-Spectrochemial analysis originated with the work of KIRCHHOFF and BUNSEN in 1860 but found
relatively little application until the 1930s. Arc-spark emission and, to a lesser extent, flame emission
methods then became popular. Following World War II flame emission became very popular. In 1955 the
modern era of atomic absorption spectroscopy began with the work of WALSH and ALKEMADE and ti-tz.
The time since 1955 can be divided into seven year periods. The first was an induction period (1955-1962)
when AA received attention from only a very few people. This was followed by a growth period (1962-1969)
when most of what we see today was developed, and then by a period of relative stability (1969-1976) when
AA contributed greatly to other fields. We are now in a period of great change, which started in about 1976,
due to the impact of computer technology on individual laboratory instruments.

hi MODERN analytical applications of atomic absorption spectroscopy trace their origin

to the two well known papers [l, 21 which were published a quarter century ago. It is
the silver anniversary of those publications that inspired this special journal issue. The
entire history of atomic absorption is much longer, however, and we will now look
briefly at a much earlier period.
All of optical spectroscopy dates from 1672 when NEWTON published his observations
on sunlight dispersed by a prism. Indeed, the word spectrum originates from Newton’s
attempts to describe the ghostly appearance of dispersed sunlight. The first to describe
spectral features other than simple color was apparently WOUASTON [3,4] who in 1802
described dark lines in the sun’s spectrum, many of which are due to atomic absorption.
WOLLASTON apparently did not attach much significance to his observations and it
remained for FRAUNHOFER, starting in 1817, to map the details of the dark lines which
bear his name. FRAUNHOFER’S work provides a good example of how a chance discovery
can change the course of a project. He was a practical glass worker and built the first
spectroscope in order to measure refractive indices precisely with the immediate goal of
producing achromatic lenses. In search of greater precision, he made observations on
sunlight with his spectroscope in a darkened room about 8 m from a narrow opening.
He noted and began to study the many dark lines, designating some of the more
prominent ones with letters, starting with “A” at the red end of the spectrum.
FRAUNHOFER showed how to measure the position of the lines precisely (achromatic
lenses now seeming less important) but attempted no explanation of their origin.
In 1859 and 1860, KIRCHHOFF, a physicist, and BUNSEN, a chemist, at the University
of Heidelberg [5,6], not only explained the origin of FRAUNHOFER’S lines but described
the analytical utility of flame emission measurements and the conditions necessary for
observation of atomic absorption. An important part of their apparatus was the burner
which BUNSEN had invented a few years earlier. For the first time spectral observations
could be made on a non-luminous, nearly transparent flame. It is interesting to note
that the general laws of emission and absorption of light, the explanation of the
Fraunhofer lines and the origin of spectrochemical analyses all depended on the lowly
burner which has been used by every chemist since BUNSEN’S time.
KIRCHHOFF’S first paper deals with general laws of emission and absorption. He
observed that bright lines which could be seen in the flame’s spectrum when a metal
[l] A. WALSH, Spectrochim. Acta 7, 108 (1955).
[2] C. T. J. -MADE and J. M. W. MZILATZ, J. Opt. Sot. Am. 45, 583 (1955).
[3] F. %, The Spectrochemical Analysis of Metals and Alloys. Chemical Publishing Company,
Brooklyn, New York (1941).
[4] M. SLAVIN, Emission Spectrochemical Analysis. Wiley-Interscience, New York (1971).
[5] G. R. KIRCHHOFF, Pogg. Ann. 109,275 (1859).
[6] G. R. KIRCHHOFF and R. W. BUNSEN, Pogg. Ann. 110, 161 (1860).
663
664 S.R. KOIRTYOHANN

salt was introduced became dark lines if a light of sufficient intensity was placed behind
the flame. His conclusions, as translated by WALSH [7] were:
“I conclude from these observations that a colored flame in whose spectrum bright
sharp lines occur so weakens rays of the color of these lines, if they pass through it,
that dark lines appear in the place of the bright ones, whenever a source of light of
sufficient intensity, in whose spectrum those lines are otherwise absent, is brought
behind the flame. I conclude further that the dark lines of the solar
spectrum . . . occur because of the presence of those elements in the glowing atmos-
phere of the sun which would produce in the spectrum of a flame bright lines in the
same position.”
He further concludes that the Fraunhofer “D” lines are due to sodium in the sun’s
atmosphere, that the “A” and “B” lines are due to potassium, and that the absence of
a Fraunhofer line corresponding to lithium emission in the flame means that there is no,
or at least not much, lithium in the sun. Origin of the Fraunhofer lines in the earth’s
atmosphere was ruled out because their intensity did not change when observations
were made at noon and at sundown in spite of the change of pathlength in the earth’s
atmosphere. Today we often see reference to the sodium “D” lines, using
FRNJNHOFER’Soriginal designation.
In a later publication, KIRCHHOFF states his law of emission and absorption [7]: “For
rays of the same wavelength at the same temperature, the ratio of the emissive power to
the absorptive power is the same for all bodies.” This deceptively simple law was to
form the basis for later theoretical developments by PLANCK on the quantum theory of
radiation and in turn to BOHR’Sexplanation of the origin of atomic spectra. The full
significance of this law to the analyst was, however, to remain unappreciated for nearly
a century.
When K~RCHHOFFdrew his conclusions regarding the composition of the sun’s
atmosphere, he had knowledge of observations published a year later with BUNSEN.The
first paper dealing specifically with spectrochemical analysis begins [7]:
“It is well known that certain substances possess the property of imparting definite
colors to flames in which they are heated. When the colored light thus produced is
analyzed by a prism, spectra exhibiting differently colored bands or lines of light are
seen. Upon the appearance of these lines of light an entirely new method of chemical
analysis can be based.”
Their light source consisted of a Bunsen burner flame into which salts could be
introduced on a platinum wire. Two telescopes were used, one for illuminating and one
for observing the dispersed light from a carbon disulfide-filled prism. The prism could
be rotated to bring the desired portion of the spectrum into view. Later versions of the
spectroscope included a scale which allowed settings to be reproduced and also
provision for illuminating the entrance slit from two sources for comparison purposes.
Their instrument contained all of the essential features of a modern spectroscope.
They experimented with flames of different temperature and with various salts of the
known alkali and alkaline earth elements. They state [7]:
“Among the great number of the salts we have examined which are suitable for
spectrum analysis in the flame, we have not met a single one which, in spite of the
great variety of the elements combined with the metals, has not exhibited the lines of
the metals. One can therefore assume that in all cases the lines of the spectrum of a
substance are entirely independent of the elements with which they are associated.”
This statement clearly established the agent responsible for the observed spectra and
justifies the popular opinion that KIRCHHOFFand BUNSENshould be credited with the
discovery of spectrochemical analysis, in spite of the fact that W. H. F. TALBOT, a
Scottish chemist, had reported some analytical observations on flames as early as 1826
[31.

[7] A. WALSH, Spectrochemistry since Kirchhofl and Bunsen, Proceedings of the Royal Austrdian Chemical
Institute 42, 297 (1975).
 A historyof atomicabsorptionspectroscopy 665

The method was not long in being put to practical application. It led to the discovery
of new elements at the rate of one per year for the first 4 years of its existence [Cs, Rb
(1860), Tl (1861), In (1864)].
In spite of the power of this new tool placed in the hands of the chemist, it was to be
many years before emission spectra were used extensively for analysis and atomic
absorp’tion was delayed even further. Spectrochemical analysis became a customary
operation in BUNSEN’Slaboratory but the use of a spectroscope for practical qualitative
analysis seems to have been neglected after his time. SLAVIN[4] writes of his own
experience:
“I remember during my course in sophomore physics in the early 1920’s our
professor’s carefully removing a brass Bunsen spectroscope and speculum-ruled
grating from the instrument cabinet and exhibiting them to the class. Such things
were for display, not for use. In the class in qualitative analysis our only contact with
spectroscopy was the flame tests for sodium, lithium, and potassium done by the
platinum wire, Bunser burner and cobalt glass technique.”
The library copy of Slavin’s book I used in preparing this history has the comment
“Still in 1971” scribed by an unknown reader in the margin beside this passage.
Reasons for the early neglect of spectrochemical analysis included extreme sensitivity
(sodium was always seen whether or not it was added), the small number of elements
which gave useable spectra from flames of the day, the complexity of arc and spark
spectra, and the inertia of the scientific community [8]. The method that would
eventually revolutionize metals analysis was, until the 193Os, largely abandoned by
chemists and left for astronomers, physicists, and others to develop. SLAVIN[4] states,
“Thus by 1920 all the conditions needed for a system of chemical analysis by
spectroscopy existed. We had excellent instruments, good photographic emulsions, a
power distribution network, and basic theory. However, chemists were very slow to
take advantage of this powerful tool, even for simple qualitative identifications. They
still relied on the classical instruments, the test tube, the blowpipe, the eye, and the
nose.”
Emission spectrochemical analysis underwent extensive growth during the two de-
cades from 1930 to 1950, based largely on electrical excitation. Flame sources were not
used for this early work because only a few elements could be satisfactorily excited. To
a surprising degree flame and arc-spark spectroscopy have separate histories. Early
attempts at flame quantitative analysis were limited by sample introduction methods.
The pneumatic nubulizer was used for that purpose by GOUY [4] as early as 1879 but
he was investigating emission processes, not quantitative analysis.
The modern era of flame photometry begins in about 1929 with the work of
LUNDEGARDH,a Swedish agronomist [7]. He used a premixed air-acetylene flame,
pneumatic nebulizer sample introduction, and photographic detection. His burner is
shown in Fig. 1 and is indistinguishable in operating principles from those used on most
AA units today. He eventually developed a completely automated system that changed
samples, controlled the exposure, developed the exposed film, and recorded a mic-
rophotometer tracing of the spectral lines. He also developed the spark-in-flame
method in which a condensed spark discharge was passed through the flame to enhance
excitation of some elements. All from a man whose primary interest was plant
physiology!
LUNDEGARDH tried direct photometry of the flame spectrum but did most of his work
photographically. JANSEN,HAYES and RICHTERimproved on direct photometry, while
SCHUHKNECHT isolated sodium and potassium lines from the flame with filters, starting
in 1937. The filter photometers used selenium barrier-layer photocells with galvanome-
ters and led to the first commercial flame instruments.
Advancements were undoubtedly slowed by World War II but filter photometers had

[8] S. R. KOIRTYOHANN and E. E. PICKEIT, in: A History of Analytical Chemishy, (Edited by H. A.

LAITINEN and G. W. EWING). American Chemical Society (1977).
666 S. R. KOIRTYOHANN

Fig. 1. Schematicrepresentationof the burner used by Lundegardb.

become sufficiently popular in Europe that MEGGERS in a 1949 review [9] wrote, “The
flame photometer appears suddenly to have come into vogue for the quantitative
determination of sodium, potassium, and calcium in almost anything.”
Flame attachments for existing spectrophotometers became available in the United
States based largely on the work of GILBERT [lo]. These instruments used a direct
injection or total consumption burner designed by GILBERT. The burner could be used
with oxygen and the higher temperatures extended emission applications to additional
elements.
The decade of the 1950s saw rapid expansion of the application of flame methods.
The advantages of using volatile organic solvents, rather than water, and of extracting
chelates of the elements of interest was developed by DEAN and others, internal
standardization in filter instruments was used by HERRMAN N in Germany, a multi-
channel flame instrument for the simultaneous determination of alkali and alkaline
earth elements in the biological materials was described by MARGOSHES and VALLEE,
the fuel-rich flame which aided in the atomization of oxide-forming elements was
developed by FASSEL, and the use of a high-temperature flame burning cyanogen and
oxygen was described by VALLEE and BARTHOLOMAY [9].
This was the state of affairs when analytical atomic absorption was first described in
1955 by two physicists, WALSH in Australia and ALKEMADE in the Netherlands. An
obvious question is, “Why 1955?” As stated earlier, KIRCHHOE had described the
conditions necessary for absorption, astronomers used absorption spectra routinely,
spectrochemists were familiar with self-absorption and self-reversal which were often
bothersome limitations in emission, and all of the necessary theory had been developed
between 1913 and about 1925. In fact, the principal theoretical reference in WALSH’S
paper is to a text published in 1934 [ll]. Hollow cathode lamps had been described by
PASCHEN in 1923 and there was one measurement, mercury, which had been routinely
done by atomic absorption starting in the late 1930s [12].
[9] W. F. MEGGERS, Annuaf Reviews, Anal. Chem. 21 (1949).
[lo] R. HERRMANN and C. T. J. ALKEMADE, Chemicd Analysis by Flame Photometry (Translated by P. T.
GILBERT). Interscience, New York (1963).
[ll] A. C. G. MITCHELL and M. W. ZEMANSKY, Resonance Radiation and Excited Atoms. Cambridge Press
(1934).
[12] T. T. WOODSON, Rev. Sci. Inst. 10,308 (1939).
 A history of atomic absorption spectroscopy 667

It might seem that atomic absorption came on the scene years, even decades, after its
time. A careful look at developments in other areas shows some reasons for the delay.
The early flame work of LUNDEGARDH was based on photographic detection. Atomic
absorption can be done photographically but it is much less convenient than recording
emission. The practical development of atomic absorption as an analytical method
could hardly have preceded the introduction of electrical detection systems. Since
hollow cathode lamps are very feeble sources and since flame emission was sometimes
troublesome, the use of photomultiplier detection and a.c. amplification were more or
less essential. The photomultiplier did not come into common use until the mid 1940s.
Thus WALSH’Sfirst application for a patent on atomic absorption in 1953 followed
without great delay the time when the needed technology was readily available.
The history of atomic absorption since 1955 can conveniently be divided into seven
year periods. The first, 1955-1962, was an induction period best characterized by what
did not happen. The years 1962-1969 were a period of explosive growth that I like to
call “The Fun Time”. This was followed by 7 years of relative stability (1969-1976) in
which AA made its greatest contribution to other research areas in need of data. A
second explosion, this time in electronics, began its impact on AA in about 1976 and
we are now in a period of change induced by the microprocessor. We will consider each
period in more detail.
When atomic absorption, the method which would eventually thrust its way into
analytical chemistry with explosive force, was demonstrated so successfully in 1955, it
caused little if any excitement in the spectroscopic community. The initial reaction of
the world was similar to that of JOHNWILLISwhen first called into the lab and informed
by an excited ALAN WALSH that he was observing atomic absorption. John is reported
to have said something profound like, “So What?” WILLIS was soon convinced,
however, and left his work in infrared spectroscopy to become one of the handful of
Australian and New Zealand workers who developed AA. MAX AMOS, JACKSULLIVAN,
D. J. DAVID, and J. E. ALLAN are other prominent names in the early AA literature.
They demonstrated the practical utility of the method time and time again only to be
largely ignored by the world’s analytical chemists. WABH traveled extensively during
this period delivering lectures and attempting to promote interest in the method. At
one point he accurately described the United States as an underdeveloped country so
far as atomic absorption was concerned. His American friends have not let him forget
that remark.
Interest among the instrument manufacturers was so low that the Australians
resorted to rather unorthodox but effective ways to obtain instruments. These efforts
eventually led to manufacture of atomic absorption spectrophotometers by Techtron,
Pty. Techtron was bought by Varian Associates in about 1967 and exists today as
Varian Techtron in Melbourne.
Slowly, first in South Africa, then England, and finally in the United States the
method began to receive attention. The advantages of the new method were rather
obvious, though often stated in exaggerated terms in the early development period. It
extended the range of flame methods to elements which had their best lines in the deep
ultraviolet region, it greatly enhanced the power of detection for numerous additional
elements, it was much less susceptible to certain types of interferences than emission
methods, it gave precision equal to the best flame emission methods (l-2%), and the
instrumentation could be inexpensive. Adam Hilger Ltd. was apparently the first of the
established instrument companies to offer commercial AA units, starting in about 1959,
followed by Perkin-Elmer in 1961.
The next 7 years were ones of explosive growth. The period opened, at least from my
perspective, with the Xth Colloquium Spectroscopicurn Intemationale, held in College
Park, Maryland in June of 1962 and closed with the 2nd International Conference on
Atomic Spectroscopy held in Sheffield, England in July 1969. The opening of this
period coincides with the beginning of my own active participation in atomic absorption
research which may explain why I think of it as “The Fun Time.” It was fun because
668 S.R. KOIRTYO~

people disagreed in almost every possible way about the best approaches to AA
measurements. Symposia at the time were spiced by frequent, usually gentlemanly,
arguments. New techniques and applications were suggested about as rapidly as most of
us could follow. Today a similar atmosphere prevails at symposia on the Inductively
Coupled Plasma. Then as now the high interest level was a heady brew for those
fortunate enough to be active in the area.
Much of what we see in AA instrumentation rests on the outcome of arguments
during the growth period. At least in certain areas of the midwestern U.S., flames
continued to emit. VELMERFASSELand co-workers at Iowa State University and ED
P~cxnrr and I at the University of Missouri, along with a few others, pointed out
repeatedly that the emission mode of operation had some distinct advantages and
should not be discarded. Today all commerical AA units have an emission capability.
The topic of interferences was sure to spark heated debate, with many arguing that
AA was much less prone than emission to chemical interferences. The very existence of
spectral interferences in AA was vehemently denied by proponents of the technique.
The state of confusion regarding interferences can perhaps best be conveyed by quoting
from a publication of the period [13].
“Atomic absorption spectroscopy does not suffer from chemical interference but
(the) presence of large amounts of anions and cations can cause pronounced effects
on absorption.”
It’s not the chemicals, you see, but those blasted anions and cations that cause problems.
The authors of the above statement should be defended because they were not writing
in their native tongue. Their’s was undoubtedly a language problem. The fact that such
a statement got past two reviewers and a journal editor is harder to understand.
A topic related to interferences was that of atomizer design. Flame emission in the
United States had developed almost exclusively with the well known total consumption
burner marketed by Beckman Instruments. The premixed burner design which domi-
nates the field today did not gain prominance without other ideas such as multiple pass
optical systems, multiple burners, and long absorption path lengths being examined and
eventually discarded. One commercial instrument was marketed with a five pass optical
system and three Beckman burners in tandem. It consumed a cylinder of hydrogen in
about an hour and sounded like a jet plane taking off. The large range of droplet
sizes injected into this type of flame was gradually shown to be the source of many
interferences (initially often confused with interferences due to operation in emission)
and the popularity of total consumption burners waned. Premixed air-acetylene flames
similar to LUNDEGARDH’S were used for analysis of about 30 elements.
Several groups attempted to burn an oxygen-acetylene flame in a premixed config-
uration. Additional elements were atomized but the flame had an unnerving tendency
to flash back (explode) due to the high burning velocity of the mixture. Atomic
absorption was extended to the refractory elements in a more satisfactory way when
AMOSand WILLIS[14] introduced the nitrous oxide-acetylene flame. This flame showed
the advantages of high temperature and strong reducing power but still had a suffi-
ciently low burning velocity to be used in burners of the Lundegardh type which had
become standard for atomic absorption by then. Elements which formed stable oxides,
aluminum and rare earths for example, were now readily measured, bringing the total
to over 60 elements which could be measured by the technique. Also certain types of
interferences which had been troublesome either disappeared or were greatly reduced in
the hotter flame.
The possibility of using atomic fluorescence for analytical purposes was first
suggested by AL~~MADE in 1962 and was studied intensively during the 1960s by
WINETFORDNER in the U.S. and WESTin England. The method has not become popular in
spite of significant advantages over atomic absorption in some applications. The reasons

[13] Y. K. CHAN and P. Y. WONG, Talanta 15,867 (1968).

[14] M. D. AMOS and J. B. WILLIS,Spectrochim. Acta 22, 1325 (1966).
 A history of atomic absorptionspectroscopy

2000-

lOOO-

1960 1965 1970 1975

YEAR

Fig. 2. World sales of atomic absorptionspectrophotometers.

are not very clear. The lack of commercial instrumentation may be part of the
explanation but more likely it is the overwhelming popularity of atomic absorption
methods. Atomic fluorescence hasn’t yet made it into the club.
The rate of growth of atomic absorption is shown by the graph of instrument sales in
Fig. 2 [15] which illustrates the success of the method better than words. In about
1966, JOHN WILLIS estimated that if the increasing growth rate continued for 20 years,
the entire land area of the Earth would be covered with AA units.
Such growth could hardly be sustained and the next period is one of relative stability.
By 1969 most of the misconceptions about AA were resolved. Spectral interferences,
while relatively rare, were acknowledged to exist and chemical interferences were
known to be primarily a function of the atomizer; i.e. the flame, not the mode of
observation. Automatic background correction systems were available to remove
residual spectral interferences. The use of releasing agents and the hotter nitrous oxide
acetylene ffame allowed rather good control over most chemical interferences. The
relative place of emission and absorption had been resolved at least in principle. In
practice, I see very few people using flame emission today even in cases where the
advantage is obvious to me. When we say that 196971976 was a stable period, let me
emphasize that we are talking about relative stability compared with the previous
period. Research was being done and improvements were being made but for one who
participated in the exciting growth period, the slower pace was quite obvious.
The arguments which had raged between the proponents of emission vs absorption
flame methods in the early 1960s had died away by 1966. The proponents of atomic
absorption had won because of vigorous promotion by instrument manufacturers (who
played an important role in popular acceptance of AA during the growth period) as
well as inherent advantages of the method. But the nitrous oxide-acetylene flame is
also a powerful emission source as was first suggested by AMOS and demonstrated by
work in the author’s laboratory. In his recent history, WALSHsays [7],
“The introduction of the nitrous oxide-acetylene flame was not only a milestone in
the development of atomic absorption; in my opinion it also represents the greatest
single advance in emission flame spectroscopy since LINDEGARDH’S original work.”
If Dr. WALSHwas thinking of improvements in the flame itself, and not in associated
apparatus, I can agree. Atomic absorption and flame emission are now regarded in
many laboratories as complementary techniques.
In 1961, L’vov [16] described his electrically heated carbon tube furnace as an
atomizer for atomic absorption. Though the original publication clearly showed high
sensitivity for the method, the design was complex and most people were too busy
exploiting the new-found capabilities of flame atomic absorption to give L’vov’s work

[15] A. WALSH, Anal. Chem. 46, 698A (1974).

[16] B. L’vov, Spectrochim. Acta 17, 761 (1961).
670 S. R. KOIRTYOHANN

the attention it deserved. WOODRIFFalso designed a rather complex furnace atomizer

[17]. A much simpler furnace was described by MASSMANN in 1968 [18] and several
modifications of his design were introduced commercially, starting in 1969. Since most
furnace work was done after the close of the rapid growth period, I have chosen to
include it with the recent developments section.
Other atomizers such as arcs, glow discharges, and hollow cathode sources have been
used for atomic absorption but have not yet had significant impact on the method.
Developments since 1969 have included most applications of fumance atomizers
including very recent and exciting work to explain and overcome many of the
interference problems encountered. Very good hollow cathode lamps are readily
prepared for most elements but a few, such as arsenic and selenium, were unsatisfac-
tory. Radio-frequency electrodeless discharge lamps, developed initially for atomic
fluorescence, have largely overcome remaining source problems. Starting with the work
of HOLAKin 1969 [19], a series of methods were developed to exploit the tendency for
some elements to form volatile hydrides. Most of the hydride formers (As, Se, Ge, Ga)
gave results which were less than satisfactory by conventional flame or furnace
methods. Hydride generation provided chemical enrichment and simplification of the
sample which were quite helpful and these methods are now routinely applied in many
laboratories. The established method of continuum source background correction has
recently received competition from a system based on the Zeeman effect developed by
KOIZUMIand co-workers in Japan and from a high resolution wavelength scanning
system developed by TOM O’HAVER’Sgroup in the U.S.
During this relatively stable period, new instrument models continued to be intro-
duced at an undiminished rate. Little was added to basic instrument functions and
many times the changes were mostly cosmetic-designed to sell instruments, not to
improve results.
Enter the microprocessor! In about 1976 the revolution in solid state electronics
began its dramatic impact on individual laboratory instruments. AA will never be the
same! But what will? The effect on our daily lives has just begun with pocket
calculators and digital watches. Who knows what the full impact will be? It would be
hazardous now to predict when this period will end or how we will be doing AA when
it does. The impact of microprocessor-controlled instrumentation is already great and
will soon be greater. Properly used, microprocessors allow us to easily do things that
are difficult or impractical without them. They can be used to assess the quality of AA
data and to enhance it with little or no operator inconvenience. They can also be used
improperly and when this is done they can lull the operator into a false sense of security
regarding his data. It will be the job of a historian a decade or more from now to decide
if we have used this new technology in the most advantageous way.
I wish to close with another brief look at the co-discoverers of AA in 1955. The
papers by WALSH and ALKEMADEwere published at virtually the same time. Yet most
AA users, if asked who invented the method, would say WALSH.Why? The answer rests
not with the original publication, but with the tireless and continuous efforts expended
by ALAN WALSH in getting a skeptical world to accept the method during the
1955-1962 induction period. He was convinced of the value of the method and had the
courage of his convictions. The honors that he has received ranging from being
knighted in 1977 to this commemorative journal issue should be attributed at least as
much to the effort in convincing the rest of us that AA was a good idea as to the
original idea itself.

[17] R. WOODRIFF and G. RAMEXOW, Spectrochim. Acta 23B, 665 (1968).

[18] H. ~~ASSMANN,Spectrochim. Acta 23B, 215 (1968).
[19] W. HOLAK,Anal. Chem. 41, 1712 (1969).