UDA5 57. RTRAFRIHSEINO ELCINO /BIOELECTROMAGNETICS EXPERINENT(U) HASHINGTON UNIV STLOUIS NO DEPT OF ELECTRICAL ENGINEERING W F PICKARD

NCASFE 9JUN 85 N8e914-94-K-0457 F/G 6/0 9

7R-iI655CI.EI O H EIN RSLCINO

414-

1 -21111

NATIONAL BUREAU Oir STANDARDSMICWOOPY RESOLUTION TEST CKART

UNCLASSI FIEDSECURITY CLASSIFICATION OF THIS PAGE (len Data Entered)

REPORT DOCUMENTATION PAGE READ INSTRUCTIONSBEFORE COMPLETING FORM

1. REPORT NUMBER 2. GOVT ACCESSION NO. 3. RECIPIENT'S CATALOG NUMBER

4. TITLE (end Subtitle) S. TYPE OF REPORT & PERIOD COVERED

CRITERIA FOR THE DESIGN OR FINAL TECHNICALSELECTION OF A BIOELECTROMAGNETICS 01/09/84 - 31/05/85EXPERIMENT 6. PERFORMING ORG. REPORT NUMBER

7. AUTHOR(s) 6. CONTRACT OR GRANT NUMBER(#)

U, William F. Pickard N00014-84-K-0457

9 PERFORMING ORGANIZATION NAME AND ADDRESS t0. PROGRAM ELEMENT. PROJECT. TASK(Washington University AREA & WORK UNIT NUMBERS

IDDepartment of. Electrical EngineeringSaint Louis, Missouri 63130

II. CONTROLLING OFFICE NAME AND ADDRESS 12. REPORT DATE

Office of Naval Research, Life Sciences Programs June 28, 1985Biological Sciences Div. (Code 441CB) 13. NUMBER OF PAGES800 North Quincy St. / Arlington, VA 22217 14

14. MONITORING AGENCY NAME & ADDRESS(II different tom Controlling Olice) 1S, SECURITY CLASS. (of this report)

Director, Office of Naval Research Detachment, Chi agoRoom 286., 536 South Clark St. UnclassifiedChicago, Illinois 60605 IS&. OECLASSIFICATION/OOWNGRAOINGSCmEOuLE

16. DISTRIBUTION STATEMENT (of this Report)

Scie ic Offic ee Blk. ,-DIIBU O.Ti - .--hn i ci!>Adm is ative . ntr cting 0 N edE .m on 2 nter,

1 1 / 1 gApproved Icx public releas-% 5 ([12] 1D" Nav Rese rch L ( IwtibUtiol Uoliwied I i St3ati9

. .hingto C.203. ria ,' A 2231417. DISTRIBUTION STATEMENT (of the abstract entered In Block 20, II dllierent Ito Report)

DTIC- I -I " FEW I i

IS. SUPPLEMENTARY NOTES t.L E TE

r- JUL 12 1985

) IS. KEY WORDS (Continue on reverse side if necessary and Identify by block number) B: ?:xperimental design)

O Nonionizing radiation" Biological effects of electromagnetic fields, .h- 20. ABSTRACT (Continu, an tovers, aide if necessaryy and identify by block number)

Z 1' 0 - ,Six criteria are propounded for judging the credibility of a reported ex-periment in athermal bioelectromagnetics. Following this, twelve criteria are

-propounded for judging the merit of the design of a reported or contemplatedexperiment in athermal bioelectromagnetics.

It is concluded that experimentation in athermal bioelectromagnetics wouldbe markedly facilitated by the identification of the mechanisms underlying theobserved effects. 0 -1 ' , c .. " -,.

DD 1jAN73 1473 ,ETION OF INOVSIISISOSOLTE UFAJ S/N 0102-LF-014-661 UNCLASSIFIED .

SECURITY CLASSIFICATION OfrTHIS PAG (When Date BtFe0l)

REPORT NUMBER: -CRITERIA FOR THE DESIGN OR SELECTION

OF ABIOELECTROMAGNETICS EXPERIMENT

Final Technical Report for the period01/09/84 - 31/05/85 onContract N0014-84-K-0457 of theOffice of Naval Research800 North Quincy StreetArlington, Virginia 22217

Submitted 06/28/85 byWilliam F. PickardDepartment of Electrical EngineeringWashington UniversitySaint Louis, Missouri 63130

CLASSIFICATION: Unclassified

DISTRIBUTION: Scientific Officer (see 81k. 11 DD2222) [1 copy]

Administrative ContractOfficer (see Blk. 8 DD2222) [1 copy]

Director, Naval Research Laboratory [1 copy]ATTN: Code 2627Washington, DC 20375

Defense Technical Information Center [12 copies]Building 5, Cameron StationAlexandria, VA 22314

14 pages

~1

SYNOPSIS

SUMMARY OF ALL WORK ACCOMPLISHED:

This is provided in the sections following.

INDEX OF ALL TECHNICAL REPORTS:

This final technical report is the only technical reportproduced on this contract.

INDEX OF ALL PUBLICATIONS:

As of the date of this submission, no manuscripts have beenformally accepted for publication. However, it isanticipated that two will eventually result; and reprintsthereof will be forwarded to the Scientific Officer whenthey become available.

LAccezrinn ForN, .

2r.

2

. .) . . . . . . .... . . . . .

PROLOGUE

The aims of this contract (NO0014-84-K-0457) were (i) toexamine the present state of bioelectromagnetics research andfrom this examination to distil a set of guidelines with whichone could more effectively design prospective experiments orevaluate extant experiments, (ii) to apply these guidelines tothe preparation of a major research proposal which focused on theneed to understand the mechanism of action of the effects of.lowlevel electromagnetic fields, and (iii) to prepare a "criteriondocument" in which these guidelines were delineated. Task (i)has been carried out; and Task (ii), which was based upon it, haslikewise been completed. This Final Technical Report is intendedto fulfil Task (iii).

In the preparation of a document intended to assist abioelectromagnetics professional in the design or evaluation ofexperimental programs, the need for concreteness might seem ofparamount importance; and the reader would therefore be justifiedin expecting great specificity of detail, with numerous realexamples of allegedly felicitous (or misbegotten) proceduresbeing provided for each guideline listed. This, however, isbioelectromagnetics, a small field in which everybody knowseverybody else; and the author is reluctant to single out in suchfashions experiments whose creators are still living. Todenigrate an experiment, however justified the grounds for doingso might seem, could conceivably provoke counterproductivecontroversy. To cite as exemplary a particular experiment could,while delighting some, leave others feeling slighted. The author _

has therefore opted for sparse footnoting and generally has leftthe reader, assumed to be knowledgeable in the field, to providehis own illustrative cases.

3

3

INTRODUCTIaN

Within bioelectromagnetics, it is usual to categorize abiological effect of an electromagnetic field of subopticalfrequencies as either "thermal" or "athermal". The terms areloose, and their use will oftentimes provoke uncertainty in someand annoyance in others. The terms are also so familiar and sowell established that to attempt to avoid them would itself becontroversial.

By "thermal" one norm lly means an effect which (i) has beenbeen explained in terms of\ the heating produced by the absorbedelectromagnetic energy or (ii) is seen primarily at levels ofabsorption where such heatinV l:ght be expected to bephysiologically significant.-y ,Gathermalf0 one normally means aneffect which %i) has been explained explicitly and unambiguouslyin terms of mechanisms other than increased random molecularmotion (i.e., heating), or rii) occurs at absorbed power levelsso low that a thermal mechanism seems unlikely, or Tiii) displaysso unexpected a dependence upon some experimental variable thatit is hard to see how heating could lie behind it.

At the present state of development of the field, ourknowledge of thermal phenomena significantly exceeds that ofathermal phenomena, but interest is skewed toward the athermal.This is not to say that thermal phenomena are no longer of anydeep scientific interest: there is still a lot to be learnedabout hyperthermia therapy for cancer or microwave assisted roomheating. Nor is it to state categorically that athermalphenomena of biological and/or environmental significance do infact exist: while many believe that they do, others would assertthat (subject to certain reservations outlined below) few if anyreported low level athermal effects have ever been successfullyconverted from the status of "claimed" to that of "wellestablished" and "well understood".

This distinction between thermal and athermal effects worksfairly well from low radio frequencies to the infrared, say from30 kHz to 30 THz (10 km to 10 jm). Below 30 kHz classicalelectrophysiological effects (including extremely low frequencybiomagnetic phenomena) can begin to appear. Althoughelectrophysiological phenomena are primarily athermal by thedefinition provided above, they are normally consideredseparately in practice; and the definition of athermal isnormally restricted to exclude them.

Thus, the term athermal refers to those phenomena whichremain when the set of all suboptical bioelectromagneticphenomena has had subtracted from it both thermal and classicalelectrophysiological effects. The principal debate inbioelectromagnetics today is between those who suspect that theathermal category is rich in subtle and unexplained phenomena

4

and those who suspect that, for practical purposes, it is a nullset. The principal question of athermal bioelectromagneticstoday is that of the mechanism(s) by which putatively athermaleffects might be produced. The principal challenge of athermalbioelectromagnetics today is that of moving reported effects fromthe category of "claimed" to the categories of "well established"and "well understood".

For an effect to be categorized as "well established" theremust be substantial agreement within the scientific communityboth about the conditions under which it can be observed andabout its characteristics when observed. It must have beensubjected to enough independent confirmation to have becomereproducible on demand; and this in turn implies (i) that it be areal phenomenon which is reproducible in principle and (ii) thatseveral groups, distinct from the one which discovered it, havecared enough and had resources enough to reproduce it inactuality. When an effect has been tied down to this extent,exploratory experiments can be superseded by experiments in thehypothesis-test paradigm and an explicit search for mechanismundertaken. Only when a mechanism for an effect has beenidentified and has achieved marked ascendency over its rivals,can the effect be said to be "well understood".

The aim of this report is to discuss guidelines by which anexperiment directed toward the discovery or elucidation of anathermal effect might be designed or evaluated. It will consistof three main sections.

First, the, great commandment of athermalbioelectromagnetics will be discussed: that any reported,putatively athermal bioeffect be robustly reproducible.

Second, six criteria for estimating the credibility of aclaimed effect will be given.

Third, twelve criteria for designing an ideal experiment willbe given.

In each case, the criterion itself will probably elicit butlittle controversy. Where controversy can arise is in theapplication of a criterion (which may not always be easy toquantify) to real life situations (which are seldom free ofconfounding factors).

5-1.

THE GREAT COtMANDMENT

Within athermal bioelectromagnetics there is a problem withonce-reported effects which either have not or can not bereproduced. The *have not" carries with it a connotation ofvolition and the sense that, if a competent investigator caredenough and had the requisite resources, it could be reproduced orat least given substantial indirect validation. The "can not"implies that respected workers have tried and failed to reproducethe effect. The latter situation is manifestly undesirablebecause its basis is not understood and could be explainedvariously:

(i) The positive findings were simply wrong. In theabsence of clearly identified blunders, this is a hardjudgement to sustain.

(ii) The positive findings were the result of experimental

artifacts. Again, unless the artifacts can beunequivocally pinpointed, this is a hard judgement tosustain.

(iii) The positive findings were due to unlikely stochasticexcursions of the experimental variables. This sourceof confusion seems especially troublesome formultiparameter exploratory studies because it is notat all unlikely that one parameter in fifty will proveto be altered significantly (p<O.05) and because theannoyances of repeating such a study can beconsiderable.

(iv) The positive findings may have required a particularconjunction of circumstances which simply are notapparent to the concerned experimenters. Phenomena inthis category are sometimes termed Cheshire-Cats.

Something obvious can be done about a reported effect in the'have not" category: a concerned governmental agency canmotivate suitable research groups to attempt replications. Thecan not" category is more troublesome because in most instances

only subset (iv) can ever be positively established, and repeatedfailures to establish it serve only to elevate the likelihood ofa non-(iv) explanation.

Because of the confusion engendered by the report of anunreproduced effect, because virtually any report of a putativelyathermal bioeffect will be greeted with a degree of scepticism,and because some claimed effects have proved evanescent upon moredetailed examination, it is imperative that any effect which oneproposes to report be robustly reproducible. All othercharacteristics of preparation and of protocol must be

-

subordinate to this. The great commandment is:Subsequent workers, following in your footsteps,must be able to aoDroximate your results withoutdifficulty!

Such reproducibility is the norm rather than the ideal ingeology, chemistry, and physics. Why not bioelectromagnetics?

Regretably, this commandment has been (perhaps even must be)honored more in the breach than in the observance. Effects seenin lab A over a period of months (or years) do mysteriously fadeaway like the Cheshire-Cat; or, despite intensivecross-consultation, an effect regularly seen in lab A may neverbe found in lab B. Certainty is elusive, and one is thereforeurged to hesitancy, reticence, and near-paranoid caution in theperformance of experiments for publication.

id"

7

.........................................

K)

SIX CRITERIA OF CREDIBILITY

Because exploratory experiments seeking athermal biologicaleffects of low level electromagnetic fields tend, more often thannot, to reveal no effects, a researcher seeking to understandathermal mechanisms would be far from ill-advised to select apreviously reported effect for his study; but, becausereproducibility can be a significant problem, the question ofprecisely which previously reported effect to select can oftenloom large. A related problem confronts the regulatory agencycharged with evaluating risk in the absence of definitive data:which reported bioelectromagnetic effects should be takenseriously and which should be ignored? There are no definitiveanswers to these questions. Nevertheless, over his fifteen yearsin the field, the author has developed half a dozen rules ofthumb for gauging the credibility of claimed effects. Inapplication these guidelines could prove intensely controversial,especially if one made his assessments public; but, as privatecalculations and in the absence of clearly superior alternatives,they may constitute a useful heuristic.

A. Does the group reporti.ng the effect have a good trackrecord? This judgement is subjective to an undesirabledegree. However, if their previous work has, in youropinion, not stood up well to the test of time, whatreason is there to presume that this work will? Isthere evidence that the group has markedly improved?

B. Has sufficient detail been given for you to proceed withcertainty in a step-by-step repetition of theexperiment? If the detail has not been given and cannot be obtained, be suspicious. Where detail islacking, the probability that you will end up chasing aCheshire-Cat increases.

C. Do the identified experimental variables admit ofprecise control, quantification, and measurement?Independent variables which cannot be well controlledmake the collection of reproducible data a virtualimpossibility. Seemingly vague nominal and ordinalscales of measurement are more prone to distortion byunconscious experimenter bias than are ratio scales forwhich accurate measurement techniques exist.

D. Are the results at variance with rational expectationbased upon previous experience? If they are, are thereplausible (even though unproven) explanations for thisin terms of well established mechanisms? Remember thatnew and unexpected biophysical principles are found butinfrequently and that most of us will never be fortunateenough to stumble upon one. Less rare, but still rare,are novel applications of familiar principles to yieldstartling results.

- 8 i

.. o... . , .. . .. . ... . . . .. _... ... .. . :-: .

E. Is the claimed effect well above the noise level of theexperiment? Effects which can be resolved only whenlarge numbers of experiments are averaged may tend to benonrobust. At the least, they are subject to maskingand distortion by normal stochastic variations in theexperimental variables. Worse, in comparison witheasily resolved effects, they are more readily affectedby unconscious experimenter bias.

F. Finally, are the individuals who reported and who workwith the claimed effect open to criticism of theirexeriments? If they are unavailable for comment, or ifthey meet reasonable criticisms with seemingly ad hocexplanations not firmly supported by experimental data,or if they appear unperturbed by the unavoidable gapswhich appear in every study, it is an inauspicious sign.

Real-world experiments are never perfect; and inflexibleapplication of these credibility criteria will seldom provepossible if one is to retain for consideration any low level,putatively athermal bioeffects. Nevertheless, zeal in asking thesix questions should serve to identify many claims which,provisionally, will fail the test of time, be relegated to thecategory of Ounconvincing evidence", and there abide with slimchance of ultimate vindication. Most claimed effects will failsome of these guidelines: the ones to avoid are the ones whosefailures are egregious.

To remove some of the abstractness from this document, thcseconsiderations will be illustrated by reference to a paper ofErich Pflomm [Arch. Klin. Chir. 166, 251-305 (1931)] in which itwas reported that an isolated frog heart, exposed to a suitableelectromagnetic field, would manifest a slowing of its beat rate.The subsequent history of this observation is complicated; but itdoes appear that fifty years of subsequent investigation have notsufficed to indentify conditions under which microwave radiationwill reliably produce bradycardia in an isolated heart [cf. K. R.Foster and W. F. Pickard, manuscript submitted to New EnglandJournal of Medicine (1985)]. The claimed effect mildly fails (A>because the Pflomm group is of unknown reputation. It grosslyfails (B> because the conditions of exposure are largelyundescribed and dose estimates are not possible. It meets -C>because heart rate is unambiguously quantifiable. If fails {D}because there is no obvious mechanism to account for the slowingand because heating by the field should produce a speeding up.It may fail <E> because the data presented in the report appearto be far from unambiguous. It presumably fails (F> because,more than fifty years after the fact, it seems unlikely that onecould engage in a deeply informative dialogue with theinvestigator. Hence it would seem a less than ideal candidate toselect for further investigation of putatively athermal effects; -

and in fact, as history shows, it violates the great commandment.

9

-:-- _.. . .. . . . . . . . ..... :? - : ._::::. . . : : . :.. _:... . . . . . .._: :-:: ": :-:: .

TWELVE CRITERIA OF IDEALITY

When a claimed effect has been judged to have an acceptableprospect of robust reproducibility, it must be asked whether theeffect is actually worth reproducing. This question has twocomponents: first, is the experiment well designed; and, second,is study of the effect apt to lead anywhere? The first componentis tricky because an answer in the negative normally leads to an"improved" experimental design which may in fact improve theeffect out of existence, either by removing an unsuspectedartifact or by inadvertently perturbing a Cheshire-Cat; and oneseldom knows which. The second component is dangerouslyprovocative but must be faced if one's interest is in themechanism of a claimed effect: the principal question relatingto athermal bioeffects is the mechanism(s) by which they mightoccur; and an experiment which reveals such an effect whileyielding no clue to its provenance is somehow unsatisfying.

No one experiment is apt to meet all the constraints whichone might place upon an ideal experiment; but an experiment whichmeets most of them is more apt to "lead somewhere" in the sensethat it will further progress toward the elucidation of amechanism. In the process of witnessing the controversies of thepast fifteen years in bioelectromagnetics, of attending uncountedconferences, and of participating in innumerable informaldiscussions, the author has identified a dozen criteria by whichhe would characterize an ideal experiment. Of these, the firsttwo relate to the electromagnetic field, the next six concern the - -biological preparation, and the final four are more "miscellaneous.

1. The fields actually present within the biologicalpreparation should be approximately known orprescribable; alternatively, the power absorbed shouldbe known approximately. Classically, this constrainthas not always been met; and the failure to meet it has -frequently been correlated with nonreproducibility.Certainly, ambiguity of dosage can only serve to inhibitthe identification of underlying mechanism.

2. There should be substantial control over the frequency,waveform, intensity, and modulation of the appliedfield. Fixed frequency experiments have long beenpredominant in bioelectromagnetics, and with goodreasons: the cost of changing frequency (or modulation)is often frightening, and the manpower requirements formultiple frequency replications of a labor intensiveexperiment no less so. Nevertheless, the unequivocal -identification of an athermal causative mechanism isbound to be difficult in the absence of an actionspectrum for the claimed effect.

10

3. The preparation should be well understood. "Wellunderstood" in this context is both subjective andrelative; but the objective of learning more about aclaimed effect is markedly facilitated by not havingfir-, to learn more about the basic biology of thepreparation. In particular, the principal investigatorhimself should be deeply versed in the natural historyand practical lore of the preparation (or personallyacquire such knowledge as a first step of theexperimental program): trusting one's subordinates toacquire and sensibly employ such information is a chancybusiness.

[At the time that the original proposal for thiscontract was written, it was envisioned that a selectlist of especially suitable preparations would beassembled from which the author at least would thenchoose his future experimental vehicles. This naivehope was never realized for two reasons. First, thenumber of preparations which are well understood in thesense that they have been the subject of at least scoresof research papers is vast: cat soleus muscle is wellunderstood, but so is the Drosophila salivary gland, orthe oat coleoptile, or the Paramecium, or the characeaninternodal cell, or the frog neuromuscular junction; andhow is one to say in advance which would prove the mostadvantageous for an exploratory study inbioelectromagnetics? Second, the author had not yetformulated the great commandment. He now believes thatany preparation which yields a robustly reproducibleathermal effect is "especially suitable" and that nopreparation, however well understood, which fails toyield such an effect is suitable: the problems ofathermal bioelectromagnetics can be resolved only by theelucidation of mechanism, and mechanism will never beelucidated in the absence of reproducibility.]

4. The preparation should be both robust and stable.Feeble preparations tend to divert the experimenter'sattention from the experiment, and finicky ones cangenerate more uncertainty than they dispel. Thenecessity of culling experiments because of

* miscellaneous vagaries of the preparation is to beavoided because it is a fertile source of unintentionalbias, especially if one's criteria for rejection oracceptance of data are at all subjective.

5. The preparation should encourage penetration to deeperlevels of understanding. Success at one level of anexperiment should bring with it a reasonable sense ofwhat steps should next be taken in a search formechanism.

6. The preparation should encourage both control and dataanalysis in real time. When days (or weeks) elapsebetween the electromagnetic exposure and the assay forthe bioeffect, it is hard to rule out the presence ofconfounding variables and to guarantee the validity of

0 11

experimental controls. When years elapse,inferences become even more difficult andpost hoc ergo Prooter hoc an ever more ulk

7. The preparation should provide biologicalof the initial transduction from electromato biological effect. Without such amplifthe most sensitive instruments may detectHowever, biological amplification may takethus there is a tension between this criteprevious one.

S. The relevant properties of the preparationprecisely and unambiguously quantifiable.against unwitting experimenter bias, and iuse of more powerful statistical tests.

9. The measuring apparatus should neither perperturbed by the applied field. Such perta much dreaded source of experimental arti

10. The design of the experiment should enableexperimenter to distinguish between effectclassical thermal or electrophysiologicalthose of more complicated provenance. Thiexceptionally important when one's stressidentification of athermal mechanisms.

11. The experiment should be emulatable. Whatother merits may be, experiments with daunrequirements of funding stability, of manpequipment, or of technique are not likelyreplicated. And, the scepticism of bioeleresearchers toward athermal effects beingunreplicated experiment is not likely to einfluence.

12. The experiment should be seminal. It shoulead somewhere. It should make a profoundthe field. A key ingredient here is thatstunning. But stunning results alone do nit is necessary also that they be engagingto a wide audience, many of whom have theacting upon them. Murky prose in an obscuwill not suffice; nor will a bizarre prepawill an arcane technique. The stunning resomehow rivet the attention of a lot of sccan put them to work - scientists who finebeset by a mountainous literature and manyprofessional pressures as well.

The experiment of Pflomm, referred to in the prsection can be cited to make these considerations wIt would appear to fail <1> rather badly since theto have been to a near zone field of presumably illcharacteristics. It clearly fails (2> for the freqfixed at about 75 MHz . It passes (3> because frcthough complicated, is a well understood organ prepis probably in some trouble at (4> since isolated Ibe tricky; however, with care, it should be possibconstraint. It passes (5> because of the wide numt

12

pharmacological challenges which can be given the preparation to

dissect out the mode of action. It excels at <6> for the effects

were alleged to have had a rapid onset. It may or may not meet

(7>: in the absence of a clearly understood mechanism, there is

no telling. It excels at <8> since heart rate is eminently

quantifiable. It may or may not meet (9): the details of the

experimental arrangement are too sparse to say. It probably

passes (10> since the expected response to heating is

tachycardia. It passes (11> for the measurements all seem easy

to make with reasonable apparatus. It fails <12> abysmally:

effects which cannot reliably be reproduced cannot reasonably be

called seminal.

13

ii4

-- " r " * * ""- . . . . . . .-vJk~.....n"*'lmim'" ll~d ml;..... m.....an...... i, ... . ....... . . . ."' . . . "-

Once a field of science has been explored well enough for itspractitioners to have a clear picture of its phenomena, researchtherein normally develops a pronounced hypothesis/test character.The goal of this development is an understanding of themechanisms behind the phenomena. And the goal of mechanism isprediction. One cannot possibly do all conceivable experiments.One cannot even do all the experiments for which there is asocietal imperative. But if one understands the underlyingmechanisms, if one has a model, then one can make some fairlyshrewd guesses of the outcome of a particular course of action;and one can limit expensive and time-consuming experimentation tothe validation of those guesses.

In athermal bioelectromagnetics today we have few wellestablished effects and fewer still agreed upon mechanisms.Because reproduciblity is not yet commonplace, virtually allexperiments are in some sense exploratory. And, statistically,historically, exploratory experiments in athermalbioelectromagnetics have small likelihood of turning up robustlyreproducible athermal effects. This has been the situation forfar too long, so long that its prolongation could conceivablydivest the field (and its practitioners) of the esteem ofscientists in less exploratory endeavors.

These problems, coupled with the limited supply of skilledmanpower in the field and the still more limited supply ofresearch funds, bode ill for any research strategy based uponunfocused exploration. At best such work will yield effects whenmechanisms are needed. The course of minimum risk for the fieldis to focus all efforts upon the replication of a very fewselected experiments: optimally, this will yield the long soughtmechanisms; more realistically, it might yield a few morerobustly replicable bioeffects; at worst, it will remove a fewadditional claimed effects from serious consideration. Failureto pursue such a course will probably lead to a perpetuation ofthe present weak correlation among the research thrusts withinathermal bioelectromagnetics and to a prolongation of thecurrently turbid status of the field. The anticipated outcome ofsuch a prolongation is the gradual diversion of this field'sfunding to other fields whose practitioners know where they aregoing.

* .The evaporation of funding for athermal bioelectromagneticswhile the present confused literature persists is not asocietally desirable outcome because it could well leave theregulators of our environment trapped between a morass ofcontradictory data and a seriously worried public; and it couldresult in a climate severely inhibitory to technological advanceswhich require even minimal "electromagnetic pollution" of thatenvironment.

14

- -- m , i ~ m ,,h;'ih~m....."

..."

... ..... ... ..... ........ "• -" ..

FILMED

8B-85

DTIC

I

I