orbits          time
             3        1978-12-07T14:30:47.000Z
           640        1980-09-05T17:14:34.016Z

          4961        1992-07-06T00:26:53.847Z
          5055        1992-10-07T19:49:39.752Z

   The source of the data and their corrections are
summarized below:

    SPECIES         M/E USED      COMMENTS
    He                  4

    N                  30         Surface recombined N and O

    O                  32         Surface recombined O to O2;
                                  corrected for CO2 fragmentation
                                  corrected for estimated surface
                                  recombination of O to CO2(*)

    N2,CO           14,28         m/e 14 corrected for NO, CO and
                                  CO2 fragmentation; m/e 28
                                  corrected for CO2 fragmentation

    CO2                44         Corrected for surface
                                  recombination of O to CO2*

    (*) the correction is based on matching scale height
        temperatures of O and CO2.

   The data are from the nonretarding potential mode of the
instrument.  Data from the retarding mode are consistent with
those obtained from the nonretarding mode and have not been
included.  The data set does not include the factor of 1.6
increase in density needed to maintain compatibility with
other data sets as discussed by Hedin et al. (1983).  Two
data sets are provided: high resolution (HIRES), every point,
composition; and low resolution (LORES), 12 second sampled,
composition.  The LORES data set represents the best
estimated composition data and is derived from the HIRES data
set.

        The field names described in RECORD 1:

     NAME      DESCRIPTION                            UNITS
     YYYY      YYYY=4 digit year
     DDD       DDD=3 digit day of year
     UT        Universal Time represented as the
               number of milliseconds since
               1966-01-01T00:00:00Z stored as a double
               precision floating point number.
     ORBIT     Orbit number
     TIMTAG    Nominal time tag assigned by project
     DHE       Number density of He                   part/cm**3
     DN        Number density of N                    part/cm**3
     DO        Number density of O                    part/cm**3
     DN2       Number density of N2                   part/cm**3
     DCO       Number density of CO                   part/cm**3
     DCO2      Number density of CO2                  part/cm**3
     DRHO      Total mass density                     g/cm**3
     DTOT      Total number density                   part/cm**3

   This data is a representative sample, approximately once per
12 seconds, of the high resolution data. It is constructed at
designated times which have been supplied by the Project.  Data
with errors greater that 30% are not included nor are data with
angles of attack greater than 40 degrees. An absolute altitude
cutoff of 250 km was used for all species except for He for
which 350 km was used.

   Each representative data point is constructed using an
exponentially weighted average of the data over a 24 second
interval centered at the sample point time.  Corrections to
the number densities of CO2 and O for surface reactions
were made at this time based on empirical model results. A
minimum of 3 data points per species and all data available
for corrections are required to be present in order for a
sample point to be output.  The total number density and
total mass density are computed if all major species (CO2,
CO, N2, and O) are present.  The data spacing is nominally 12
seconds except for the -12, 0, 12 time tags. Although time
tags from -1800 to 1800 seconds are generated, only those data
records for which at least one species has a valid value for
that time tag are output.

   No spacecraft positional parameters have been included
in the data sets.  These can be obtained from the SEDR data
submitted separately by the Project.

Several criteria were invoked when inserting data for a given
orbit: orbit and attitude parameters must exist (project supplied);
the spacecraft format and bit rate must be appropriate for
acquisition of data by the ONMS; and the command sequence for the
instrument must be appropriate for useful determination of
atmospheric composition. Cases where useful composition cannot be
determined include special test modes (e.g., retarding potential
sweeps, filament off) and 1/8 unit amu sweep modes.  In addition,
composition for the LORES data set cannot be easily determined for
unit amu sweep mode.  The ONMS was not operational for every orbit
nor is every orbit complete due to data gaps introduced by use of
telemetry formats for which the ONMS has no instrument output.

  Useful composition data are gathered from the lowest periapsis
altitude to a maximum altitude generally around 250 km (about 300
km for He).  The actual maximum altitude depends on the accumulated
surface gas buildup acquired from previous orbits which creates a
gas background.  The gas background was estimated from high altitude
averages of the data and for all species, except helium, an inbound
signal/background ratio of 2 and an outbound signal/background ratio
of 4 were used as cutoff values.  In some cases superthermal ions
(e.g., Kasprzak et al.,1982) were observed at low altitudes (e.g.,
below 300 km for orbit 219) and these were removed when visually
detected.  Some problems have been observed in the high altitude
data very near cutoff, particularly for outbound N2. Several data
points were never removed and appear higher than the expected
extrapolation of the data to that time.

   Residual spin modulation which had not been completely removed
is evident in the processed data. The source of the spin residuals
are the gas/surface adsorption/desorption effects which were not
removed from the data and a noncosine behavior for the response of
the ion source density with angle of attack.  Another feature
observed occasionally at large angles of attack (>40 degrees) is a
reduction of the data when compared to data at lower angles of
attack. This has been determined to be due to antenna shadowing;
that is, the ONMS geometric view cone 'sees' the spacecraft antenna
at extreme angles of attack. Occasionally near minimum angle of
attack (<10 degrees), enhanced data points are observed for m/e=4
(He channel) which are apparently high energy ions/neutrals
traveling along the tube axis and being detected.  The more extreme
points in either of these two cases have been mass flagged.

   The data time spacing depends on the spacecraft bit rate and
format, and the particular instrument commands executed.  Usually
programmed mass format was used but occasionally unit amu and 1/8
amu sweeps were implemented.  Several orbits switched from low
electron energy to high electron energy and as a result there may
be a discontinuity at the transition point.  The 1/8 amu sweep data
have not been included.

   Atomic nitrogen was measured in programmed mass mode only
after orbit 190.

   Orbits 1-19 generally do not have reliable relative composition
due to the fact that gas-surface processes in the ion source had not
stabilized. This affects all surface reactive species except He.

   Isolated (one or two points per several spin cycles) high
resolution data points are occasionally observed and they should be
regarded as erroneous points which are more likely wrong than right.

   The error associated with the points is more an indication of data
quality than of absolute uncertainty.  It contains the statistical
error of the data determined for the principle m/e used for the species
from the detector signal plus the errors coming from any other species
used to correct the data. It also contains a contribution which is
proportional to the background/signal ratio.  The total relative error
is at least an additional 5-10% above this value.