A REMOTE ACTION INVESTIGATION WITH MARINE ANIMALS

• Document Type: CREST
• Collection: STARGATE
• Document Number (FOIA) /ESDN (CREST): CIA-RDP96-00787R000300070001-3
• Release Decision: RIFPUB
• Original Classification: U
• Document Page Count: 48
• Document Creation Date: November 4, 2016
• Document Release Date: June 11, 1998
• Sequence Number: 1
• Content Type: REPORT

Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
NOMil `id
-"101111,SP,..-
Final Report?Objective E, Task 1
A REMOTE ACTION INVESTIGATION WITH
MARINE ANIMALS
By: EDWIN C. MAY
C. M. PLEASS
College of Marine Studies
University of Delaware
Prepared for:
PETER J. McNELIS, DSW
CONTRACTING OFFICER'S TECHNICAL REPRESENTATIVE
333 Ravenswood Avenue
Menlo Park, California 94025 U.S.A.
[phone]
Cable: SRI INTL MPK
TWX: 910-373-2046
For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
December 1987
215
I;=
For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Final Report?Objective E, Task 1
Covering the Period 1 October 1986 to 30 September 1987
I I
1:C:32=5
(SdA REMOTE ACTION INVESTIGATION WITH
MARINE ANIMALS
EU=D
By: EDWIN C. MAY
C. M. PLEASS
College of Marine Studies
University of Delaware
Prepared for:
PETER J. McNELIS, DSW
CONTRACTING OFF ICER'S TECHNICAL REPRESENTATIVE
SRI Project 1291
Approved by:
MURRAY J. BARON, Director
Geoscience and Engineering Center
333 Ravenswood Avenue ? Menlo Park, California 94025 ? U.S.A.
[phone] ? Cable: SRI INTL MPK ? TWX: 910-373-2046
d For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
December 1987
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
ABSTRACT
In FY 1986, SRI International awarded a subcontract to the College of Marine Studies of
the University of Delaware to conduct remote action (RA) experiments using marine algae as
target elements. Protocols were developed during that year that would enable SRI to test, with a
living system, the Intuitive Data Sorting model. During FY 1987, significant improvement was
made to stabilize the data so that standard analysis techniques (e.g., ANOVA) might be used.
While much progress was made toward that end, significant autocorrelations persist.
Nonetheless, an attempt was made to generate successful RA. SRI analyzed data for four
participants and found no statistical evidence of RA.
ii
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
TABLE OF CONTENTS
ABSTRACT
LIST OF FIGURES
ii
iv
INTRODUCTION
1
II REVIEW OF THE UNIVERSITY OF DELAWARE REMOTE
ACTION PROJECT
2
A. Toward Increased Data Stability
2
B. Protocol for Data Collection and Analysis
3
C. Results
5
D. Discussion
5
REFERENCES
8
APPENDIX
A-1
iii
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
LIST OF FIGURES
1. Results of the IDS and Monte Carlo Calculations 4
2. Post Session Control Velocity Spectrum--Operator 48 6
3. Post Session Control Velocity Data--Operator 48 6
iv
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
I INTRODUCTION
For a number of years, Dr. C. M. Pleass of the College of Marine Studies at the University
of Delaware, has been constructing a remote action (RA) experiment using various species of
marine algae as targets. In these types of experiments, RA participants attempt to change the
swimming velocities of single algae cells during effort periods as compared to control periods.
Encouraged by a paper presented at an annual meeting of the Parapsychological
Association,1* SRI International awarded a two?year contract to the University of Delaware to
satisfy Objective E, Task 1 of the FY 1986 statement of work. For the first year, the task was to
stabilize the equipment and protocol, while the task for the second year was to collect RA data.
This report describes SRI's role in that effort and summarizes the work done by Dr. Pleass.t
(The University of Delaware's draft final report is included, verbatim, in the Appendix.)
* References may be found at the end of this report.
t This report constitutes the deliverable for Ojective E, Task 1.
1
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
II REVIEW OF THE UNIVERSITY OF DELAWARE
REMOTE ACTION PROJECT
In the past, the putative RA data from the College of Marine Studies of the University of
Delaware, have been criticized for the statistical methods that were employed. The cell velocity,
as a function of time, exhibited an unknown mixture of biological and environmental correlated
responses (i.e., the data points were not statistically independent). In a report included in SRI's
FY 1986 interim report on this experiment,2 a significant autocorrelation was observed, yet there
was no modification of the statistical analysis to account for the correlation. (The complete FY
1986 results can be found in Reference 2.)
During the first few months of FY 1987, the University of Delaware moved the project to
new and better facilities. The move delayed our joint FY 1987 project: however, as is shown
below, a major improvement of data stability was realized.
Initially, the experimental setup in the new location was identical to the old, but with the
following improvements:
? The floor of the new laboratory is a 3-foot thick concrete slab. This improved
the physical isolation of the apparatus.
? The apparatus was installed in an electrically isolated and sound attenuated
room.
? It was now possible to collect data for 24-hour periods.
SRI personnel inspected the new facilities and found them to be significantly improved. What
follows is a description of the FY 1987 activity.
A. Toward Increased Data Stability
Baseline velocity data were collected over 24 hours for fifteen 24-hour periods. It was
clear (see Figure 2 in the Appendix) that the cells exhibited erratic behavior during the "day"
(0800 to 1700) but were quiescent between 0000 and 0800. This quiescent period suggested that
the organisms might be sensitive to the increased electromagnetic activity during the day and
might be less stimulated during the night. (SRI analyzed a quiescent data set and found the data
to be normally distributed with a slight tail favoring downward swimming.)
2
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
A second possible contribution to the data instability was noticed. The algae cells tended
to clump, especially if they had been in the experimental chamber for some time. If this
tendency was consistent, it would produce strong autocorrelations in the data.
The solution to the first problem was to attempt to provide better shielding, but it was
realized that ELF shielding was not possible. Instead, a completely new experimental chamber
was designed.
In the old system, cells, which were drawn once per day from a biostat, constituted a fresh
experimental sample daily. While it was good to have fresh cells, the immediate change in
environment led to initial instabilities and, in the absence of light and nutrients, the cells lost
energy over time. This trend was obvious from the data in that cells showed an increasing
tendency to swim with gravity rather than against it.
The new system uses a continuous flow design. Fresh cells flow through an experimental
region at an orthogonal (to the measurement axis) velocity of ?450 mm/sec. This change greatly
improved the stability, and significantly reduced the clumping.
B. Protocol for Data Collection and Analysis
With the increased stability, it was decided to use two types of analysis:
(1) Standard Analysis of Variance (ANOVA).
(2) Data Analysis sensitive to the Intuitive Data Sorting (IDS) model.
The standard ANOVA 2 x n consists of a design where n is the number of consecutive temporal
data samples collected in each of the effort and control conditions. The IDS analysis was
outlined earlier,2 and a complete description can be found in SRI's final report on Objective E,
Tasks 3 and 4.3
Briefly, the IDS formalism uses a simple application of the central limit theorem to show
that for mean chance expectation (MCE), the expected value of the log of the absolute value of
the detrended velocities is linearly related to the log of an averaging length, n. Under RA
conditions, the relationship is nonlinear in a dramatic way (Figure 1). Under IDS conditions
(i.e., psychoenergetically biasing a sampling distribution) the formalism predicts a linear
relationship once again, but with increased intercepts over MCE. The averaging length, n, is
defined as the number of samples of the parent velocity distribution that are averaged to form
one data point.
3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
_ed
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Change in Velocity (Nersalixel)
Monte Carlo
Theory
Note: Only a few points
are shown for clarity.
J000000:0,
RA (a=.0.30)
IDS (1=0.38)
MCE
10
Averaging Length (n)
SO
100
600
FIGURE 1 RESULTS OF THE IDS AND MONTE CARLO CALCULATIONS
SRI conducted extensive Monte Carlo calculations to simulate the experimental IDS
design. The results are shown in Figure 1. The RA curve was calculated assuming a parent
distribution shift of 0.3a--the only free parameter. The IDS curve was calculated assuming that
the sampling distribution had an increased variance2 of 0.38. Only 4 of the 12 generated data
points (3 for each model) are shown. All points fell within 1a (the error bars) of the theoretical
curves--an expected result considering that each model has only one specified free parameter.
SRI proposed that 50 trials for each of three averaging lengths (5, 15, and 40) be collected
by one RA operator (blind to the averaging length).
Delaware proposed that the actual data collection proceed as much like their "wave"
protocol as possible. In a single trial, an operator marks a continuous, but undisplayed velocity
data stream when he/she feels that RA has just occurred. This effort period is followed
immediately by a control period of length n points. This is followed by a brief, but random,
interval before a second identical period to the above couplet is collected as a pseudo-effort
period.
In addition to the psychoenergetic trials, SRI requested that a "real" influence be exerted
upon the algae under the same collection format described above. The data would be used to
4
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
demonstrate that if RA behaved like a "real" influence, the IDS formalism would be capable of
describing it.
C. Results
Four operators produced five trials each at each of the three averaging lengths. This raw
data and one set of "real" influence (spark stimulus) data simulated for five trials was sent to SRI
for analysis.
The "wave" protocol used by the the Delaware group suggested that the ANOVA be
changed to a 2 x 2n design. A separate analysis was then performed for each of the averaging
lengths. The first row of the ANOVA data matrix consisted of the five-trial averages of the
velocity (detrended) for each of n points during the effort period (i.e., just before the operator
marks the data stream) and a second set of n control points. The second row was identical to the
first row, except the data were taken from the pseudo-effort period.
This particular design was sensitive to the "release-of-effort" wave protocol and to
differences between the effort and pseudo-effort periods. The ANOVA was conducted for all
three averaging lengths for the four operators and for the spark stimulus data. There was no
evidence even suggestive of an extra-chance anomaly.
Delaware made two attempts at simulating RA. The first used spark induced E&M as a
perturbing force, however, the intensity was insufficient to produce any effect. Following an SRI
suggestion, Delaware used visible light to induce a change in swimming velocity. SRI was unable
to confirm a stable velocity shift in the data from this attempt as well.
D. Discussion
Two conditions must be met before an ANOVA is valid. The data must be approximately
normal and they must be statistically independent. ANOVA is particularly forgiving for the first
condition, but if a positive serial correlation exists, a significant underestimate of the residual
variance can result.4 This can result in an artifactually significant F ratio.
Figure 2 shows a typical velocity distribution with a Gaussian fit and a Gaussian fit with a
matched exponential tail. It is clear by inspection and the value of chi-square that the first
criterion (approximate normality) is easily met.
The second criterion is much more important from an ANOVA perspective. While the
autocorrelation has improved greatly since the earlier report,2 it nonetheless, still represents a
5
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
high degree of significance--especially for the early lags. This can be seen in Figure 3 as low
frequency oscillations of the velocity envelope. Even after the significant improvement of the
system, some non?random component to the velocity data remains. Because the overall
ANOVA result was not significant, and the autocorrelations are all positive, we are not required
to calculate the appropriate error residuals for our correlated data.
Telocity lenity (irbitrary)
?er.
e Data Points
? Gaussian Fit
? ? Gaussian + Tail
?40 ?20 ?20 ?10 0 10 90 SO 40
Velocity (micrometers/sec)
FIGURE 2 POST SESSION CONTROL VELOCITY SPECTRUM -- OPERATOR 48
-
-
...... ....
.11
-:
Z:
I
I ill 1
I
I
I
I I
?
I I I
I
0 1000 0000 8000 4000 11000
Time (-sec)
FIGURE 3 POST SESSION CONTROL VELOCITY DATA -- OPERATOR 48
6
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
The Delaware efforts were quite disappointing. SRI designed the protocol and analysis
and made numerous suggestions about improving the physical set-up. Virtually all the RA data
were collected in the last month of the contract, so that only 5 trials instead of the requested 50
were collected at each averaging length. Second, the operators were not blind to the averaging
length; thus, even if we had observed significant evidence of an anomaly, we would be unable to
interpret it with regard to the IDS model.
Why the RA simulations (spark data) were not significant remains a mystery. Either the
wrong stimuli were chosen by Delaware, or the stimulus intensities were incorrect.
7
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
REFERENCES
1. Pleass, C. M., and Day, N. D., "Using the Doppler Effect to Study Behavioral Responses
of Motile Algae to Psi Stimulus," Proceedings of the 28th Parapsychological Association
Convention, pp. 373-406 (August 1985).
2. May, E. C., Humphrey, B. S., and Pleass, C. M., "Measuring Remote Action Influence
on the Vertical Component of Dunaliella Velocity," Interim Report, Objective E, Task 9,
Project 1291, SRI International, Menlo Park, California (December 1986).
3. May, E. C., "Intuitive Data Sorting: An Informational Model of Psychoenergetic
Functioning," Final Report, Objective E, Tasks 3 and 4, Project 1291, SRI International,
Menlo Park, California (December 1986).
4. Neter, J., Wasserman, W., and Kutner, M. H., "Applied Linear Statistical Models,"
Second Edition, Homewood, Illinois, Richard D. Irwin, Inc., p. 445 (1985).
8
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
APPENDIX
A-1
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
1
Final Report on the Stanford Research International Project.
"Evaluation of the Theory of Intuitive Data Sorting, and Its
Application to Bio Laser Doppler Data Generated by Marine
Microorganisms"
Duration: Two Years from 10/1/85
Program Leader:
Dr. C. M. Pleass
College of Marine Studies
University of Delaware
Newark, Delaware 19716
The following description is excerpted from the original proposal,
so that this final report may be placed in proper perspective:
Year One - 10/1/85 - 9/30/86._
Using the existing Bio Laser Doppler facilities at the Bayside
Laboratory, develop a protocol and statistical analytical procedure
which can effectively address the question of the validity of the
SRI theory known as Intuitive Data Sorting (IDS).
The deliverable, due by 9/30/86, would be a report describing
a mutually acceptable experimental and analytical technique. This
implies frequent contact with the SRI project monitor throughout
the year, as the various options are examined.
Year Two - 10/1/86 - 9/30/87.
Using the protocol and procedure developed in Year 1, to
carry out Bio Laser Doppler experiments that address the validity
of the IDS theory. This will require the involvement of volunteer
participants from the Lewes community, whose task will be to try to
Intuit or anticipate changes in the statistical parameters
describing the motion of marine microorganisms, or in some similar
SIRPROJ
CMP/vs
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
2
way, to provide evidence bearing on the validity of the SRI theory.
The individuals involved will not be identified in the report by
name, and the appropriate informed consent will be obtained in
each case, to meet the University of Delaware's human subject
requirements.
The deliverables will be a report containing a description
of the detail of each trial, including statistical ensembles that
best represent the data,, and copies of the floppy discs carrying
the raw data.
In practice, the project progressed in two cycles, both of.
which involved program development and experimental psi work. Thus
in Year 1, two appropriate psi protocols were developed (pk 85 and
WAVE), and volunteers from the Lewes community used them to create
data bases which SRI and UD examined together'. The experimental
and analytical techniques appeared acceptable, a priori, but
instability in the baseline time series velocity data threatened
the credibility of the results.
Very few segments of our time series records of unstimulated
algal movement taken prior to the fall of 1986 were free from some
abrupt changes in average velocity and vector. However for
practical reasons most of the data had been acquired during the
1 Algal swimming velocity, resolved along a single axis,
constituted the primary experimental variable. The variance of
this data and the direction of movement, up or down, were secondary
variables.
SIRPROJ
CMP/vs
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
3
working day. This suggested 24-hour time series studies of
unstimulated cultures, to examine the possibility that the observed
changes were part of a natural circadian rhythm. After modifying
the system to allow 24 hour studies, fifteen 24-hour data sets
were collected in November - December 1986. On January 13, 1987,
Ed May and Peter McNelis of SRI met with Mic Pleass and Dean Dey
at Lewes to discuss this data. A typical time series is shown in
the fold-out (Fig. 1).
A "window" was clearly'evident in all 15 data sets between
circa midnight and eight AM. (Fig. 2) The data in this region
were quiet and would have formed an acceptable basis for psi.
experiments. However, it seemed unlikely that participants and
experimenters would show much enthusiasm for working through this
period.
As a secondary benefit, the data gave us perspective on the
hours 0900 - 1500 which include most of our previous psi runs. It
seemed clear that this was a period when the algae were generally
swimming non-randomly, with quite frequent abrupt changes in average
velocity. As we were all aware, this places a heavy stress on the
analytical technique and the credibility associated with the data
from psi experiments. It is impossible to choose between exogenous
or endogenous (circadian) influences when looking for a cause.
However, the group were suspicious of sonic and electromagnetic
disturbances associated with the working day.
SIRPROJ
CMP/vs
.0 Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
24-HOUR DATA
TEST850
24-HOUR DATA
TEST850A
24-HOUR DATA
TEST850B
24-HOUR DATA
TEST850C
16:33:42 ?17:20:30 18:08:07 18:58:01 19:4.3:42 20:39:41 21:28:38 22:16:5:2 23:0c;:31
TIME
24-HOUR DATA
TEST850D
TIME
24-HOUR DATA
TEST850E
3 : 54:15 00:42:58
TIME
241-1OUR DATA
11 EST850F
01:3::00 02:24:06 03:16:41 04:10:29 05:04:42 05:57:14
TINE
24-HOUR DATA
TEST8500
-100
-150-
-200-
-25n
05:57:14 06:48:00 07:36:35 08:24:31
09:12:1? 09:59:51 10:,]8:08 11:35:43 12:23:54 13:11:41 13:59:18 14:4;7:06 15:35:05 16:22:49
TINE
TIME
TIME
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
? ? '.;?.?
? ill I I I il I I I 111 1 I I I II I I MI I I I I I I I I I I ",
------- ---- ---- r ? ? ,
-t
--#1.414444"i' .1
:-.011110.1irligile
:-.1,000110.011011010111
r'fki?lOWORMOMMUS
41/
;' ,'?
;.`. .:??
i. g.
1...1 'f..;?????1'4'1.--:;;CL.,--;;,:::
. .... ' ?
''...':-..;,'?i ..?
i; 'IL ? .
, ..., .,
;?-,, '''.
p ?::::',.:015- -
s. ? ?????? ...:1?.`31.!???a, ? ,? ?.1
?
'M00%
ti
a??:"...
?Ii
y '?
: # t w 'te," 4 4.4?%
-
?
Fig. 2 TIME SERIES VELOCITY DATA
? Vertical lines are at
midnight and 0800 hours
R-7"Pc rr/c-r-larr.,71^,rTr?r?
?!,
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
6
The environmental chamber housing the BLD apparatus was
converted into a modestly good RF and sonic quiet room. The results
were unchanged. The next step was to actively look for evidence
of group dynamic change within the algae in the tissue culture
vessel used to hold the sample. By visually observing a particu-
larly hardy clone of Dunaliella which lived for three months on
Mb c Pleass' .office window sill (unnurtured through the period, and
knocked over several \times!) we were able to identify "micro
avalanches" of cells, superimposed on upward and downward movement
of the green suspension. Literature search revealed one relevant
reference2 which provided examples of microavalanching in cultures'
of Dunaliella. This seems to be a natural method of redistributing
cells which have physical reasons for normally swimming upward.
To work around this problem a flow system was constructed,
so that a well mixed culture from a biostat external to the
environmental chamber could move in slow (100-200 um sec-') laminar
flow past the measuring volume. The virtual fringes of the BLD
system were aligned parallel to the flow, ie, the axis along which
velocity measurements were made was chosen to be at right angles
to the artificially induced flow. Experiments confirmed that these
velocity data were unaffected by the slow flow. Twenty-four hour
runs appeared much quieter. Ed May visited Lewes again on May
5th, just at the right time to see a very good looking time series
2 Kessler, J. O., Cooperative and concentrative phenomena of
swimming microorganisms. Contemp. Phys. 1985, Vol. 26, #2, 147-166.
SIRPROJ
CMP/vs
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
0
-fi()
-40
-60
-80
-100
-120
-140
-160
-180
-200
-:Z20
-240
Fig. 3
24 hour velocity data from Dunaliella
24- HOU R DATA
FLOW TEST 004
.24- HOU R DATA 21,
FLOW TEST 004 0
-
-
...,
--,
_
^
-
^
1
II
,
Release 2000/08/10 : CIA-RDP96-007871
^
_
-
-
- - ,
A
2 4 6 a 10 12 14
TIME OF DAY
18 18 20 22 2 4 6 6 10 12
TIME or LAY
cect, Sjci, )&61,t61/4;:-__. 1 Ack,i-rk
4.12_11. 1QtAdvv,v,-,- 0,v /0 V-ovt----s.
44,;(g- c)t, G.o...f,C 4_
14
"C)
OQ
????41
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
Nuttier S tikA
pcmvAA
Variable:
',.E LO
VELI
tIEL2
Sample size
1658
1757
1890
Average
-162.039
-162.306
-162.547
Median
-162.312
-162.312
-162.312
Mode
-161.547
-163.078
-163.078
Geometric mean
Variance
39.0124
37.563
36.1619
Standard deviation
6.24599
6.12886
6.01348
Standard error
0.153394
0.146216
0.138323
Minimum
-185.281
-184.515
-182.219
Maximum
-127.859'
-137.813
-141.641
Range
57.4219
46.7029
40.5785
Lower quartile
-166.14
-166.14
-166.907
Upper quartile
-157.719
-158.484
-158.484
Interquartile range
8.4217
7.6563
8.4226
Skewness
0.190437
9.197301
-0.0216271
Standardized skewness
3.16568
3.37628
-0.383844
Kurtosis
0.553595
0.325558
-0.0216545
Standardized kurtosis
4.60128
2.78553
-0.192164
Variable:
VEL3
VEL4
VEL5
Sample size
1857
1926
2096
Average
-163.34
-163.527
-163.456
Median
-163.078
-163.844
-163.844
Mode
-162.312
-161.547
-163.844
Qeometric mean
Variance
38.1946
41.748
37.9794
Standard deviation
6.18018
6.46127
6.16274
Standard error
0.143415
0.147228
0.13461
Minimum
-196
-189.109
-181.453
Maximum
-139.343
-140.875
-139.343
Range
56.6564
48.2338
42.1094
Lower quartile
-167.672
-167.672
-167.672
Upper quartile
-159.25
-159.25 I
-159.25
Interquartile range
8.4217
8.4217
8.4217
Skewness
0.112429
0.0642093
0.167328
Standardized skewness
1.97791
1.50873
3.12743
Kurtosis
0.561965
0.316777
0.0616489
Standardized kurtosis
4.94322
2.83777
0.576123
Fig. 3A
Summary statistics of one hour periods
from data set 004
8
72tcooci
(9o:o0
o6:
dweit 4,.
/4.- ?
citc.0v3 &AI owa;
*ta.nraf4..,S
Sele.a.4 v..4.411
1.4.,A?id...0
caisf.
cuct,
azkitt.
t3CACr
0
("kiaz.-1-,LaL
offvv-
dx". or,14,4
t-e? 04_ lis
17.4A4rt. erce.):44c).-
k co-1
OtitYkrs. Liza jr
'
at.44. 104. tEt
ti,rtz_41.3h4s Akre
LAJ444
Cc.u..#04czksm,01.4.4
444k- it,s
Art.
.41c
Approved For Release 2000/08/10 : CIA-RDP96-00787R000300070001-3
I