IEEE TRANSACTIONS ON INDUSTRY AND GENERAL APPLICATIONS, VOL. IGA-4, NO.
5, SEPTEMBER/OCTOBER 1968 467
Reevaluation of Lethal Electric
CHARLES F. DALZIEL, FELLOW, IEEE, AND
Abstract-Low-frequency electric currents of a few milliamperes
flowing through the body cause muscular contractions. In the arm
such an effect may make a subject unable to let go of a live conduc-
tor. The highest currents which 99.5 per cent of men and 99.5 per-
cent of women are able to let go have been shown to be 9 and 6
mA, respectively. Currents somewhat larger than this, in the range
of 20 to 40 mA, passing across the chest may arrest respiration lead-
ing to asphyxia, unconsciousness, and even death.
The most common cause of death in electric shock probably is
ventricular fibrillation, a condition in which the circulation is ar-
rested and death ensues very rapidly. An analysis of available experi-
mental data indicates that body weight and shock duration are im-
portant factors in determining the maximum current not likely to
cause ventricular fibrillation. Taking a weight of 50 kg as the average
for a human victim it is suggested that the relationship between cur-
rent and shock duration is given by I = 116/V/T, where I is the
current in milliamperes and T is the time in seconds. It must be
stressed that this has only been shown to be valid within the range
of 8 ms to 5 seconds.
Currents flowing through the nerve centers controlling respiration
may cause respiratory inhibition, which sometimes persists for a
long time after the current has been interrupted. Other effects pro-
duced by high currents, such as burning, etc., are not discussed in
this paper.
INTRODUCTION
THERE IS increasing evidence that the most common
mechanism of death from electric shock is ventricular
fibrillation, and it is therefore of considerable practical im-
portance to establish the maximum values of commercial
frequency alternating currents that are unlikely to cause
fibrillation. As the circumstances of an electrical accident
only rarely allow a reliable estimate of the current or of the
period for which it flows, it is necessary to have recourse to
animal experiments to determine threshold currents. This
method introduces difficulties of extrapolation of the find-
ings to human beings. Until recently the only two series of
animal experiments of any magnitude in this field have
been the studies of Ferris and his colleagues in 1936 at
Columbia University and the Bell Telephone Laboratories,
N. Y. [1], and of Kouwenhoven and his associates at The
Johns Hopkins University, Md., in 1959 [2]. The results of
these studies were analyzed by Dalziel in 1960 [3 ], with the
object of establishing an acceptable relationship between
the three factors believed to be concerned, namely, body
weight, current magnitude, and shock duration.
Paper SL 68-1 approved by the Group Safety Liaison Committee
for presentation at the 1968 IEEE Industry and General Applications
Group Annual Meeting, Chicago, Ill., September 29-October 3.
Manuscript received August 6, 1968.
C. F. Dalziel is with the Department of Electrical Engineering,
University of California, Berkeley, Calif.
W. R. Lee is with the Department of Occupational Health, Uni-
versity of Manchester, Manchester, England.
Currents
W. R. LEE
Since that time there have been two further develop-
ments. First, Kiselev of the U.S.S.R. Academy of Medical
Sciences published in 1963 a detailed experimental study
of threshold currents in dogs [4], and second, in 1966, Lee
published detailed studies of low-voltage fatal electric acci-
dents in the United Kingdom [5 ] and questioned the analy-
sis used by Dalziel [6].
This paper presents a brief review of electric stimulation
of nerves and muscles as observed in studies of let-go cur-
rents [7 ], [8 ]. The fibrillation data from the various sources
will then be compared. First, the relationship between cur-
rent and shock duration will be examined. Next, the rela-
tionship between body weight and fibrillating current when
the duration of the shock is held constant will be studied.
Finally, since several of the studies reported fibrillation in
terms of the minimum fibrillating current, the relationship
between minimum fibrillating current and the maximum
nonfibrillating current will be examined in the attempt to
establish more accurately the likely fibrillating threshold
current for man.
DISCUSSION
Electric Stimulation of Muscles
Except for its nuisance value, or for its startle effect in
causing an involuntary movement triggering an accident,
the smallest electric shock of importance is that current
which causes a loss of voluntary control of the hand when
grasping an electrified object. Upon increasing the magni-
tude of the alternating current, the first sensations of
tingling give way to contraction of the muscles. The
muscular contractions and accompanying sensations of
heat increase as the current is increased, and finally a value
of current is reached at which the subject cannot release
his grasp of the conductor. The maximum current a person
can tolerate when holding an electrode and still let go of
the conductor using muscles directly stimulated by that
current is called the let-go current. Let-go currents are im-
portant, as an individual can withstand, with no ill after-
eff ects, repeated exposure to his let-go current for at least
the time required for him to release the conductor. Cur-
rents only slightly in excess of one's let-go current are said
to "freeze" the victim to the circuit. Such currents are
very painful, frightening, and hard to endure even for a
short time. Points representing let-go currents for 124 men
,and 28 women are shown in Fig. 1 [7], [8]. In these tests
the subjects held and then let go of a small copper wire,
and the circuit was completed by holding the other hand on
a flat brass plate, or by clamping a conducting band
wrapped with saline soaked cloth on the upper arm. The
effect of size of the electrodes was also investigated. It was
468 IE4E TRANSACTIONS ON INDUSTRY AND GENERAL APPLICATIONS, SEPTEMBER/OCTOBER 1968

L/2 )9.5-
ti -
99-
998 ----1 98 _ _ _
2z y
z 95-
95
Er
90 -- 90 Z _ -
w 80 AVERAGE FOR w
-i - - _ -J 80 -
F::L 28 WOMEN-* F:
z 70 lOSma z 70 _
w w - -
0 60 _ _
MEN - 16 moFOR 134
AVERAGE
= j _
a.) 50 50 - - ~~~r Ir
w 50-
40- ----
a. 40 = =
w 30 30
a- - --

cr 20 - - - - - -
w 20 _
Q (D-
w
0 10 I!) z
10
w 5 6mo --
w
C3) 5 -
C.) wr 2 I

_IXi23LI ('a %).>-nma4

9
0.5 l i -- 0.5 I4-l(1/2 %)650 ma _

6 8 10 12 14 16 18 20 22 24 400 1000 2000 3000 4000

LET-GO CURRENT MILLIAMPERES (RMS) FIBRILLATING CURRENT MILLIAMPERES (RMS)
Fig. 1. Let-go current distribution curve for men and women, Fig. 3. Minimum fibrillating current distribution curve for dogs,
60-Hz commercial alternating current. 8.3-ms shocks, I1/2% = 650 mA, 150o% = 2070 mA, Iav = 2070
mA.

919.5 _ l
99.
98 ..-
:
4 95 _
z

850 0/| :
90

HInIFnI 1HnT
InITI1 InrHII IrTTl
80 - - -- - --

7A p z
LI
a:
70
6
--

U,
LIda.0
rId 60 c -
20 - l I I
it 50 VY 4

408
10

i /T
-

z
ui
hi6
a:2

IN
PERCENTILE CURVES
P.: a. 2
LLI
I. I (I/z%0) :BOO0ma ___I
0
400 1000 2000 3000 4000
1~ FIBRILLATING CURRENT MILLIAMPERES (RMS)
hi
Fig. 4. Minimum fibrillating current distribution curve for dogs,
16.7-ms shocks, 11/2% = 800 mA, I50% = 1450 mA, Ia, = 1500
A~~~ ~ ~ ~~~~~~~5 mA.
DSAfE
CURRENT X RT
vI
The average let-go current was approximately 16 mA for
II0 50 100 500 1000 5000 men and 10.5 mA for women.
FREQUENCY HERTZ
Although average values are interesting, it is important
Fig. 2. Effect of frequency on let-go current for men. to determine the minimum likely let-go current for the gen-
eral population. After much consideration, it was decided
to specify as a reasonably safe limit for electric shock the
found that the location of the indifferent electrode, the 1/2 percentile value. This procedure is justified by the
moisture conditions at points of contact, and the size of the observation that in all cases involving a considerable num-
electrodes had no appreciable effect on an individual's let- ber of points, the actual values near the 1/2 percentiles are
go current. The individual points representing the maxi- always greater than the theoretical values. This is inevi-
mum value for each subject are plotted in ascending order table because of the inherent characteristics of the mathe-
to show percentile rank as a function of current. The data matics of probability, as it is evident that there exists a
for the 134 male subjects and the 28 females follow a finite probability that even zero current would theoret-
straight line on probability graph paper and are therefore ically freeze a certain number of men to an energized wire.
normally distributed. As would be expected, the larger the As a result of these experiments it is generally accepted
number of points for a given test, the closer the points fell that the safe let-go currents are 9 and 6 mA for men and
to the straight line which represents the distribution curve. women, respectively.
DALZIEL AND LEE: LETHAL ELECTRIC CURRENTS 469

9'9.5 - =-I- 11
9E 99--
9 99-- 98
98--
z 95 _ _ _ _ _
DISTRIBUTION CURVE
--I dt- z

4r
95
990-9DISTRIBUTION CURVE
2
8 00
-J 80 - _- - - -- -J

z 7 0
z
llJ SJ

I (50%) 1800 ma
__0 -I(!42%)-240ma | |
a.
80 -_ / - |2 ~ 0D
L&J
2C-. 0r
(2 C5 _ _ l l
zSLI
z a-
LJ
LIJ U

I(VP/%) = 240 ma 10 - 20500 00 10

a. 2----____. ooo%1500

0 - 1000 2000 3000 4000

FIBRILLATING CURRENT MILLIAMPERES (RMS) FIBRILLATING CURRENT MILLIAMPERES (RMS)

Fig. 5. Miinimum fibrillating currenit distribution curve for dogs, Fig. 7. Minimum fibrillating current distribution curve for dogs,
83.3-ms shocks, II/2% = 240 mA, I5o% = 1800 mA, Iay = 2040 0.333-second shocks, I/2% = 200 mA, I5o% = 740 mA, lav = 775
mA. mA.

99 - - - - l - -- - l
98,5- 1 98 - --------

z z 95 - - ~ .1- -- - L

DISTRIBUTION CURVE 4r DISTRIBUTION CURVE

90 90--- - 1~-
90 -
LiJ
-i 80 ---. ----

z
7C -- w
z 70 --.
UL
C-)
C - I(50%)= 900 ma ___ 6C
-0 7-
m50%)=
oI
( 15 0 ma
40 - t9
a-
40C----
4C

:
20-- lLI 20- -
(2 (2

IC ---
z z

5-
211
LI
I
I
a-
I I Ab -T---]
I l/ %)1=50 ma
0- . I- , - -

0.5--
0 1000 2000 3000 0 100 200 300
FIBRILLATING CURRENT MILLIAMPERES (RMS) FIBRILLATING CURRENT MILLIAMPERES (RMS)

Fig. 6. Minimum fibrillating current distribution curve for dogs, Fig. 8. Minimum fibrillating current distribution curve for dogs,
0.167-second shocks, I1/2% = 550 mA, 150% = 900 mA, Iav = 1240 1.0-second shocks, I1/2% 50 mA, Io0% 150 mA, Iav
= 169 = =

mA. mA.

Fig. 2 shows results of tests made to determine the effect
of frequency on let-go currents. Sinusoidal currents from 5
to 10 000 Hz were used in these experiments. There is
essentially ino differenice in the let-go current values for
frequeincies between 50 and 60 Hz which are the commercial
frequencies used throughout the world.
The muscular reactions caused by commercial frequency
alternating currents in the upper ranges of let-go currents,
typically 18 to 22 mA or more, flowing across the chest
stopped breathing during the period the current flowed,
and in several instances caused temporary paralysis of the
-middle finger. However, normal respiration resumed upon
interruption of the current, and no adverse after effects
were produced as a result of not breathing for short
periods.
In 1961, Lee described 30 cases of persons who were
"frozen" to low voltage circuits for varying periods but
who subsequently became free and survived. In these cases
the currents clearly were above the range of let-go currenits,
but they were less than the fibrillating current. However,
victims who suffered the longest shocks showed signs of
impending asphyxia during the period they were held on
to the circuit [9].
The following observations based on let-go experiments
conducted over a 25-year period may be of value in inter-
preting fibrillating currents to be discussed in the next
sections.
1) An individual's let-go current is essentially constant,
aind is not depressed by repeated let-go trials provided an
interval sufficient to permit recovery from fatigue is
allowed before the next shock is administered (actually let-
go currents taken at weekly intervals increased slightly
with time, probably due to lessening of fear, understand-
ably associated with the experience) [8].
470 IEEE TRANSACTIONS ON INDUSTRY AND GENERAL APPLICATIONS, SEPTEMBER/OCTOBER 1968

Il -1I -
98 ---5300-
98
--
iiEzzziII 2
z 95--
z 95 cr 5100-
DISTRIBUTION CURVE 4 90-
9C
-i 80 -
-j 80 - z 70 -

z 70 -- -
600
w 60 L
a:Q- I (50 %) 50 ma aI.
50

Id Sc- _ 40
0- 40 - 30-
30 -- 20
0-----
2 CD
4
(1
z
10
z
U]
0 5
a: a. 2-
U
a- 2
I'z)50ma
0.5 --I-0-2%) 1400
a-----
0.5 J
0 100 200 300 400
400 1000 2000 3000 4000
FIBRILLATING CURRENT MILLIAMPERES (RMS) FIBRILLATING CURRENT MILLIAMPERES (RMS)

Fig. 9. Minimum fibrillating current distribution curve for dogs, Fig. 11. Minimum fibrillating current distribution curve for sheep,
2.0-second shocks, 11/2% = 50 mA, I50% = 150 mA, Ia. = 187 30-ms shocks, I1/2% = 1400 mA, I50% = 2480 mA, Iav = 2630
mA. mA, Wav = 59.9 kg.

99.5- --
9!19.5
99-
1990
98 1 -
9
z
4 95- z 95
4
90 - -
10,300-~
Id 90
IL
-J 80-._- w 80
F
-i
z 70- 70
---
z -.-

60 w
Id
--

cr
,
50 II
- w
a. 40 a. 602 - - -

30 1I i5
20 - w
CD CD
-

z z A 1_ 1
Ir
5 2- w
0
w
a. a.

4 (. '/2%c/)- 25 ma I ( '^/o)-
Q 600 ma1T1< I1 f1
05Li
-

C 100 200 400 1000 2000 3000 4000

FIBRILLATING CURRENT MILLIAMPERES (RMS) FIBRILLATING CURRENT MILLIAMPERES (RMS)
Fig. 10. Minimum fibrillating current distribution curve for dogs, Fig. 12. Minimum fibrillating cLirrent distribution curve for sheep,
5.0-second shocks, 11/2% = 25 mA, 150% = 83 mA, Iav = 83 0.10-second shocks, 11/2% = 600 mA, 150%o = 2200 mA, Ia,, =

mA. 3310 mA, Way = 73 kg.

2) An individual can tolerate, with no adverse effects, two factors, fibrillating current and shock duration, is
repeated exposure to the reactions associated with currents independent of body weight. In fact, this assumption will
of his let-go level. be shown to be justified by the findings at the end of this
3) The physiological reactions leading to the inability section.
to let go are essentially the same over the limited fre- The Kouwenhoven experiments covered shock durations
quency range of 50 to 60 Hz. from 8.3 ms to seconds, and the dog was used as the
experimental animal. The distribution curves are given in
Figs. 3-10. The Ferris experiments spanned shock dura-
Relationship Between Fibrillating Current tions from 30 ms to 3 seconds and were made on several
and Shock Dur-ation of the larger animals, comparable in body weight to man,
As noted in the introduction, the three factors believed and included sheep, dogs, calves, and pigs. The distribu-
to be concerned are body weight, current magnitude, and tion curves for sheep are given in Figs. 11-16.
shock duration. It seems reasonable to assume that, within Fig. 17 is the distribution curve for the Ferris dogs, and
the range of body weight of the large animals considered, Fig. 18 is a similar distribution curve for the Kiselev dogs.
dogs, sheep, calves, and pigs, the relationship between the The Ferris series included 10 dogs ranging in body weight
DALZIEL AND LEE: LETHAL ELECTRIC CURRENTS 471

99
'9.5-ii1I I 9 19.5--
99

98 --
11400 98 ----
y
z 11300 -
z 95 - _ _ r
r
90 1100__
9900Z _
w
-J B0. - _A 9200A
_ w
-J 90
z 70 - - 7000-~ z
w _ __ _ _ __ i-
-.
6C
---- w
80
50 500 I100Sy
a-
3C -- -- w
CD
w 20 --- F-
tL

z z
FIBRLLATNG CRREN MILIAMPRES RM0
Lii w
a. a. 2 -

0.5 I,-(12'/%)= 400 mc Q~Iii A-(%) 350 ma__-

400 1000 2000 3000 4000 0 500 1000 - 500
FlBRILLATING CURRENT MILLIAMPERES (RMS) FIBRILLATING CURRENT MILLIAMPERES (RMS)

Fig. 13. Minimum fibrillating current distribution curve for sheep, Fig. 15. IMinimum fibrillating current distribution curve for sheep
0.12-second shocks, II/2% = 400 mA, I5 o = 3400 mA, Iav = 4860 0.47-second shocks, second series, I1J2% = 350 mA, '5o% = 1080
mA, Wa. = 61 kg. mA, Iav = 1080 mA, Way = 64 kg.

9 ~9.599 -I--f F
- -----
99
98 - --

98 y
z 95
z 95 2-50
90 -- --

90 - w
w -j ----_
-j 70 -

z
z 70 ---- 60
w 60 C-) -- --

-- S, w 50
a-
--

w 50 - 40 --

a- 40-
30 -- --

30 cr 20 --

20 / ID
H
9 z
CD C-)
LL
w -
a.
a. -
I2/%) 120mM-
0.511 1 [- 100 200 300 400
C 500 1000- 1500
FIBRILLATING CURRENT MILLIAMPERES (RMS) FIBRILLATING CURRENT MILLIAMPERES (RMS)

Fig. 14. Minimum fibrillating current distribution curve for sheep, Fig. 16. Minimum fibrillating current distribution curve for sheep,
L 0.47-second shocks, first series, Iav = 930 mA, Way = 61 kg. 3.00-second shocks, II/2% 120 mA, 150% = 240 mA, Iay
= =

253 mA, Way 57.4 kg. =

from 18 to 26.7 kg, who received 3-second 60-Hz shocks
while Kiselev's investigation used 35 dogs with body
weights varying from 5 to 24 kg who received 50-Hz shocks
of the same duration. Thus the Ferris results were obtained
with heavier dogs and the Kiselev results were obtained
with a larger number of dogs, over a greater range of
weights, but on the average they were lighter. It is believed
that these data are comparable and may be combined to
obtain a larger number of points upon which to base
statistical conclusions. The test procedure in these in-
vestigations was to apply a series of well-spaced shocks of
gradually increasing intensity until fibrillation occurred. In
each case the electrodes were attached to the right front
limb and the opposite rear limb, and involved a current
pathway diagonally across the chest (the common current
pathway in many fatal human accidents). In all of these
tests the shocks were applied to include the phase of the
heart cycle sensitive to fibrillation, and the effects of the
small difference between the frequency of the shock cur-
rents is believed unimportant. Fig. 19 is the distribution
curve obtained by combining the 45 results for 3-second
shocks on dogs obtained by Ferris [1 ] and Kiselev [4].
With the exception of the Kouwenhoven data for very
short shock durations (see Figs. 3 and 4) in which normal
distributions were obtained, and for the Ferris data for
0.47-second shocks to sheep (Fig. 14) which if extrapolated
would imply the untenable conclusion that 1 to 3 percent
of the animals would fibrillate on application of a shock of
zero current, the other responses are quite similar, and are
believed typical of ventricular fibrillation. In each case the
472 IEEE TRANSACTIONS ON INDUSTRY AND GENERAL APPLICATIONS, SEPTEMBER/OCTOBER 1968

99.5
99
98
2
Z 95 _ ____ ___ -
_____ -215-
cr-
90
i 80
F 70
60
, 50
L 40
30
, 20

H
z
10
<) v

2
0
0.51 ____L-I ('Y)50ma ___ __
0 50
_LJIIEL_
100 150
FIBRILLATING CURRENT MILLIAMPERES (RMS)
Fig. 17. Minimum fibrillating current distribution curve for dogs
(Ferris), 3.00-second shocks, 11/2% = 50 mA, Iso% = 90 mA, Iav =
106 mA, Wav = 22.3 kg.

Id'9.5 - - - - - - -J -
99
98
- - - -- - -- 99
995 0
98-
=_ 0 _ 0 0 2115-
X
2z
95 l l l z 95-
cr
90 -
w
90 -- --
-J 70~~~~~~~~ -J 80 _ _ _ i
z
F-
z 70 --1
w 60 - 60 _ - - 1
w 50 50 r
L
----

0- 40 - 40
30 - -- 30
20 wU
ct: 20
w a.
(2
z wL
z
w LI)
C) 5-
w
0-
di-
o51 I- V- I((/ %)~ 30 md 0.5 - MO ('/2 )32
- -

0
I50 I 100 150 c 50 100 150
FIBRILLATING CURRENT MILLIAMPERES (RM S) FIBRILLATING CURRENT MILLIAMPERES (RMS)
Fig. 18. Minimum fibrillating current distribution cturve for dogs Fig. 19. Minimum fibrillating current distribution curve for 35
(Kiselev), 3.00-second shocks, I1/2% 30 mA, I50% 70 mA, = =
Russian and 10 American dogs, 3.00-second shocks, I1/2% = 32 mA,
Ia. = 74 mA, Way = 12.0 kg. 150% = 75 mA, Ia. = 81.3 mA, Way = 14.3 kg.

response is a straight line on probability paper from low Fig. 20 shows the 1/2 percentile points and the lowest
values to at least percentile 50, and is then skewed at the experimental points for the various tests on dogs taken
higher values. The somewhat unusual response is attrib- from Figs. 3-10, and from Fig. 19 when plotted on log-
uted to the complicated nature of ventricular fibrillation; log graph paper. The Kouwenhoven points total 191 with
the exact phenomenon remains unknown, but is possibly shock durations ranging from 1/2 cycle to 5 seconds. The
due to two or more modes of excitability of the heart Ferris-Kiselev combined data resulted in the point at 3
muscle. seconds and represents 45 points. The straight line repre-
Since the responses consistently follow a normal dis- senting the response was drawn by eye as governed by the
tribution below percentile 50, the analysis leans heavily 1/2 percentiles, and is below all observed experimental
upon evaluating the 1/2 percentile points. The average, or points. The line has a slope of -1/2, and may be repre-
mean value or percentile 50, so useful in most statistical sented by an equation of the form I = K/V-T mA.
analyses, is of limited importance here because of the A similar analysis using the data from Figs. 11-16 for
skewed nature of the response at the higher values, and the sheep covering a shock duration from 0.03 to 3.00 seconds
1/2 percentile must be obtained extrapolating the data is given in Fig. 21. The minimum fibrillating points totalled
considerably beyond the limits of the experimental data. 99, and here one point (from Fig. 14) falls below the line
DALZIEL AND LEE: LETHAL ELECTRIC CURRENTS 473

The average minimum fibrillating currents for 3-second

iT,
shocks and the average body weights of animals compa-
U1) Uo
rable to man in body weight, consisting of 45 dogs, 25 sheep,
.L
-Er
.w
CD
0
a
11 calves, and 9 pigs, are presented in Fig. 23. The regres-
- x
cf) sion line for the dogs has been plotted on this figure and
< Z 100o
8 x LuJ extended to cover the range for the larger animals. Note
that except for pigs, the line passes close to the average
J x IL.

3: +
values for all of the other animals. From this it is con-
Z °
cr LuJ cluded that the minimum current required to produce
3 0 loo ventricular fibrillation is approximately proportional to
<D x body weight, not only within the single species, dog, but
z o
z X among the larger animals, probably including man. Possibly
I( 1/2 NO) POINTS
-i
*
X LOWEST EXPERIMENTAL POINTS the reason the average value for the nine pigs is much lower
than the regression line is due to the very small number of
10 pigs available, and to the special anatomical feature of the
2 animal, namely the relatively greater weight in the legs as
0.01
SHOCK
0.1
DURATION
1.0
SECONDS compared to the other animals. Obviously the configuration
of the four legs does not materially affect the current dis-
Fig. 20. Relation of minimum fibrillating current to shock duration tribution in the region of the heart.
for dogs.

Relationship Between Minimum Fibrillating and

10000 Maximum Nonfibrillating Current
The 1/2 percentile points representing the minimum
fibrillating currents for the larger animals are shown in the
x central part of Fig. 23. Due to the manner of conducting
< 1000 the tests, which consisted of applying serial shocks until
fibrillation occurred, the actual minimum current just
required to produce fibrillation must have been somewhat
zw
less than that given by the 1/2 percentile line. In 1940 a
'Jaw
member of the Ferris team supplied the senior author with
CC100 a copy of their original laboratory data including body
weights, minimum fibrillating currents, and the penulti-
mate values, i.e., the value of current immediately pre-
z

F '/2 %) POINTS
x LOWEST EXPERIMENTAL POINTS ceding fibrillation which is herein called the maximum
nonfibrillating current. Figs. 26 to 29 give the maximum
10
nonfibrillating current distribution
E~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Io I curves for dogs, sheep,
IL

Z 0.01 0.1 1.0 0

calves, pigs. The 1/2 percentile points from these maximum
nonfibrillating current distribution curves are plotted in
1

2 SHOCK DURATION SECONDS

Fig. 21. Relation of minimum fibrillating current to shock duration Fig. 23 to fix the lowest line on the figure. The actual
for sheep. fibrillating current was certainly above this line. From the
above it is obvious that the likely current just necessary
to produce fibrillation for a given mammal is somewhere
representing the locus of the 1/2 percentile points, but between these two 1/2 percentile lines, depending upon
again the line representing the response has a slope of body weight.
- 1/2. Recently Lee made a detailed case analysis of elec-
trocutions which occurred in England and Wales during
Relationship Between Fibrillating Current and Body Weight the years 1962-1963 [5]. Of these, 166 occurred on voltages
In Fig. 22 are plotted the results of 3-second shocks from less than 250 volts 50 Hz, and 30 percent were females
the 35 Kiselev dogs and the 10 Ferris dogs. The straight and 26 percent were persons under 20 years of age. It would
line representing the line of best fit, the regression line, was therefore seem prudent to assume body weight for the
calculated by the method of least squares. Although there typical victim of electrocution at less than 70 kg, the
is some scattering about this line, the trend of the response average body weight for an adult male, previously used
clearly shows that the minimum current required to pro- by Dalziel [3 ]. There is certainly a case for selecting a most
duce fibrillation is approximately proportional to the conservative value for the body weight for use in evaluating
individual's body weight, and the degree of association, K for the electrocution equation for man, that is, a healthy
indicated by the correlation coefficients, is r +0.74. normally developed human adult, and it is proposed that
=

The minimum fibrillating current distribution curves for 50 kg be accepted as a conventional value for the typical
calves and pigs are given in Figs. 24 and 25. victim of electric shock.
474 IEEE TRANSACTIONS ON INDUSTRY AND GENERAL APPLICATIONS, SEPTEMBER/OCTOBER 1968

200 8 qq 1111111
IV) 98 _ ____ _ _
cr
z 95 - - ---470-0
CD
0 0
0
E
150
w im~ w
-J 80 _ - _ _
z
w
tn
-J cn
8
<
m
w
z
C-)
70
U, tr
cr w 50 = _ = = = = =
w
0-
ooo
0
z
LO ) 0 w 20 _ _ _
-J
-J z
-i Sr 0
w
C-)
U.

:2
S *I
. * G e 0 w
a. 2
_ - _ v _=__ = =-_
1LO
151
ml
50 12
z

* Kiselev dogs
0. _- I
11____
100 200 300 400
0 Ferris dogs FIBRILLATING CURRENT MILLIAMPERES (RMS)
Fig. 24. Minimum fibrillating current distribution curve for
calves, 3.00-second shocks, I'/2% = 150 mA, 150% = 300 mA,
0 10 20 30 Iav = 312 mA, Wav = 70 kg.
BODY WEIGHT kg

Fig. 22. Relation of minimLumn fibrillating current to weight for dogs,

3.00-second shocks.
99.5
99
98
y
w
Z 95
90
*Denotes average /

values - _ 80
iii i
Iav= 3.68W+28.5 F 70
2 300 w

__
L_ - -
/

60
- 1.1 1 11
w

n /

r 50 I(%12m
cn C\
X 40 _ IDILIEIII
a- a/0I CL 30
, 20
00
z
5

200[
(I)
0D cr-
0
a w
2
z
I1/2% MINIMUM
cr
'i
:2
0
o3
x
w
FIBRILLATING
CURRENT LINE 0.51
CD) -J0 0 100 200 300
z
FIBRILLATING CURRENT MILLIAMPERES (RMS)
I /15 OX hAAVIAlI LA
., 100
-J Fig. 25. Minimum fibrillating cturrent distribution curve for
CD pigs, 3.00-second shocks, I1/2% = 120 mA, I5o% = 235 mA, Ia. =
67 ma 235 mA, Wav = 78.7 kg.

and
0 20 40 60 80 100
I 116 to 185 5 seconds
BODY WEIGHT kg rms mA.
T J8.3ms
Fig. 23. Relation of fibrillating current to body weight for various
animals, 3.00-second shocks. The question naturally arises as to the applicability of
the electrocution equation beyond the time limits upon
which it is based. The shortest time for which data are
The constant K for the electrocution equation for man available represents a half wave of 60-Hz ac, and the
is obtained by entering the abscissa of Fig. 23 at 50 kg, majority of shorter shocks might be classified as impulse
and proceeding vertically to the two 1/2 percentile lines: shocks; see Dalziel [10]. There is little information about
the effects of shocks of longer duration than 3 to 5 seconds.
I = KI /T
From 5 seconds to 20 or 30 seconds the threshold may re-
K= VTI main fairly steady, dropping only slightly [4]. For longer
K = V/3 .X 107 = 185 maximum periods, there is some evidence that asphyxial changes may
increasingly exert their influence and lower the threshold
= X 67 = 116minimum even further [11].
DALZIEL AND LEE: LETHAL ELECTRIC CURRENTS 475

9)9.5 ---- 9.~9.5

sceK
i 98 ---

9
-

z
z 95 -._
_ _ _ 199-' 95-

ct- 90 _ _ = -

G
9070~~~~~~~~~
__ =_ _ -a
* - w 80--
-J
C,
ci:z 70-

--
z
w 60- 60 l
c-Q 50-
w . 0~w 50 - --- --
a. 40 -----
8° -0 0 - -0 I- 40
w 30
30 --

20- 20
II _ _

U
w
CD,
10~~~~~~
0
CR
z
C-) IL
cQ
co:
a. 2
0- 05 __ -zl 1 H 1 I
0.5 1112]IJIIZIII7 0.5
O I('k%/)=9Omoj
50 100 150 100 2.00 300 400
FIBRILLATING CURRENT MILLIAMPERES (RMS) FIBRILLATING CURRENT MILLIAMPERES (RMS)
Fig. 26. Maximum nonfibrillating current distribution curve for Fig. 28. Maximum nonfibrillating current distributioni curve for
dogs, 3.00-second shocks, 11/2% = 35 mA, I50% = 80 mA, Iav!= calves, 3.00-second shocks, I1/2% 90 mA, 15o% = 260 mA, =

91.8 mA, Wav = 22.3 kg. Iav = 268 mA, Wav = 70 kg.

99.5 19.5..
-

99 99
98 98.----
2z 2z
95 95-.
x
90
Er
90-.
W 80 -i
80~~~~~~~~~~
z
70 z 70-_ _

i
60
50
60-.
X 40 a. 40 - .
30 301 T7-
I II 1 1
20 20-.
T- 10
z
IL
C-)
5
LLI
C- 5-

1(/IQ2%)-=7TSr 11II 1
I 1T r ------
a.

0.5
- -
J
r T
A I0
(°)=60)
6rn m
AI
-
100 200 300 400 0 100 200 300
FIBRILLATING CURRENT MILLIAMPERES (RMS) FIBRILLATING CURRENT MILLIAMPERES (RMS)

Fig. 27. Maximum nonfibrillating current distribution curve for Fig. 29. Maximum nonfibrillating current distribution curve for
sheep, 3.00-second shocks, I1/2% = 75 mA, I50% = 220 mA, Iav = pigs, 3.00-second shocks, I'/2% = 60 mA, 15o% = 210 mA, Iay =
229 mA, Wv = 57.4 kg. 198 mA, Wa. = 78.7 kg.

It is important that ventricular fibrillation follows a nor- LETHAL EFFECTS OF ELECTRIC SHOCK
mal distribution at low values of current with the prob- 1) If long continued, currents in excess of one's let-go
current may produce collapse, unconsciousness, asphyxia,
ability of fibrillation increasing up to about percentile 50,
and then the probability decreases as indicated by the 'and death.
skewed nature of the distribution curves at the higher cur- 2) Ventricular fibrillation is probably the most common
rents. At higher currents fibrillation does not occur, and cause of death in electric shock cases, and may be produced
very high short shocks stop a fibrillating heart. This is the
by moderately small currents which cause over stImulation
basis for defibrillating techniques, which are now standardof the heart rather than physical damage to that organ.
hospital operating room routine for stopping fibrillation, When fibrillation occurs the rhythmic pumping actionl of
which frequently results from several causes in addition tothe heart ceases and death rapidly follows.
electric shock. 3) Shocks administered to hundreds of anlimals iindicate
Currents considerably in excess of those just necessary that the minimum commercial frequency electric current
to produce ventricular fibrillation may cause cardiac causing ventricular fibrillation is proportional to body
arrest, respiratory inhibition, irreversible damage to the weight and inversely proportional to the square root of
nervous system and serious burns. However, no numerical shock duration. For a current pathway between major
data are available regarding current magnitudes necessary extremities in 50-kg mammals, the relationiship is approxi-
for these effects. mately (116 to 185)/\/T mA. Itis believedthat ventricular
476 IEEE TRANSACTIONS ON INDUSTRY AND GENERAL APPLICATIONS, SEPTEMBER/OCTOBER 1968

fibrillation in a normal adult worker is unlikely if the shock
intensity is less than 116/VT mA, where T is in seconds.
4) Currents flowing through the nerve centers con-
trolling breathing may produce respiratory inhibition,
which may last for a considerable period even after inter-
ruption of the current.
5) Cardiac arrest may be caused by relatively high
currents flowing in the region of the heart.
6) Relatively high currents may produce fatal damage
to the central nervous system.
7) Relatively high currents may produce deep burns,
and currents sufficient to raise body temperature sub-
stantially produce immediate death.
8) Delayed death may be due to serious burns or other
complications.
REFERENCES
[1] L. P. Ferris, B. G. King, P. W. Spence, and H. B. Williams,
"Effect of electric shock on the heart," AIEE Trans., vol. 55,
pp. 498-515, May 1936.
[2] W. B. Kouwenhoven, G. G. Knickerbocker, R. W. Chestnut,
W. R. Milnor, and D. J. Sass, "Ac shocks of varying parameters
affecting the heart," AIEE Trans. (Communication and Elec-
tronics), vol. 78, pp. 163-169, May 1959.
[3] C. F. Dalziel, "Threshold 60-cycle fibrillating currents," AIEE
Trans. (Power Apparatus and Systems), vol. 79, pp. 667-673,
October 1960.
[4] A. P. Kiselev, "Threshold values of safe current at mains fre-
quency," in Probl. of Elec. Equipment, Elec. Supply, and Elec.
Measurements (in Russian), Sb. MIIT, vol. 171, pp.47-58, 1963.
[5] W. R. Lee, "Deaths from electric shock in 1962 and 1963,"
Brit. med. J., vol. ii, pp. 616-619, September 1965.
[6] -, "Death from electric shock," Proc. IEE (London), vol.
113, pp. 144-148, January 1966.
[7] C. F. Dalziel, E. Ogden, and C. E. Abbott, "Effect of frequency
on let-go currents," AIEE Trans., vol. 62, pp. 745-750, Decem-
ber 1943.
[8] C. F. Dalziel, J. B. Lagen, and J. L. Thurston, "Electric shock,"
AIEE Trans., vol. 60, pp. 1073-1079, December 1941.
[9] W. R. Lee, "A clinical study of electrical accidents," Brit. J.
industr. Med., vol. 18, pp. 260-269, 1961.
[10] C. F. Dalziel, "A study of the hazards of impulse currents,"
AIEE Trans. (Power Apparatus and Systems), vol. 72, pp. 1032-
1043, October 1953.
[11] P. H. Gerst, W. H. Fleming, and J. R. Malone, "Increased sus-
ceptibility of the heart to ventricular fibrillation during meta-
bolic acidosis," Circulation Res., vol. 19, pp. 63-70, July
1966.

Charles F. Dalziel (M'33-SM'39-F'57) was born in San Francisco, Calif., on June 6, 1904. He
received the B.S. degree with honors in 1927, the M.S. degree in 1934, and the E.E. degree in
1935 from the University of California, Berkeley.
He was relief substation operator, Midland Counties Public Service Corporation, Santa
Maria, Calif., in 1923. From 1927 to 1929 he was testman and student engineer with the
General Electric Company, Schenectady, N.Y. He was with the San Diego Gas and Electric
Company, San Diego, Calif., from 1929 to 1932 in charge of system protection. From 1929 to
1932 he taught at San Diego State Teachers' College, San Diego, as an Instructor in the Ex-
tension Department. Since 1932 he has been with the Department of Electrical Engineering,
University of California, and in 1968 became Professor Emeritus. He was Supervisor, En-
gineering Science Management War Training Program, University of California, 1941 to
1944, and Chief Technical Aide, National Defense Research Committee, Office Science Re-
search and Development, New York, N.Y., from 1944 to 1945. He was Senior Member of a
review board for U.S. Army engineers in Tokyo, Japan, in 1951. During 1951-1952 he was
Fulbright Visiting Professor to the Istituto Elettrotecnico Nazionale Galileo Ferraris, Turin,
Italy. He was appointed Commissioner of Examinations (electrical), California State Board
of Registration for Civil and Professional Engineers, for 1954-1956. He has served as a Consultant to Kaiser Engineers, Oak-
land, Calif., in 1955; Los Alamos Scientific Laboratory, Los Alamos, N. Mex., from 1960 to 1965; Lawrence Radiation Labora-
tories, Berkeley and Livermore, Calif., since 1962; and the Rucker Company, Oakland, since 1965. He was the U.S. Dele-
gate to the Meeting of Experts on Electrical Accidents and Related Matters, International Labor Office, Geneva, Switzerland,
October 1961.
Prof. Dalziel is a member of Eta Kappa Nu, Tau Beta Pi, Sigma Xi, and is a Registered Professional Engineer in the State of
California.

W. R. Lee was born in London, England, in 1922. He received the M.D. degree from Guy's Hospital Medical School, Uni-
versity of London, London. He also holds the D.I.H. degree.
He was Squadron Leader, Royal Air Force Medical Branch, and Medical Officer, British Railways. He is presently Reader
in Occupational Health, University of Manchester, Manchester, England, Editor of the British Journal of Industrial Medicine,
and Medical Adviser, North Western Electricity Board. His special field of research is death from electric shock.
Dr. Lee is a fellow of the Royal Society of Medicine and a member of UNIPEDE Medical Group and the Society of Occu-
pational Medicine.