Analysis of REDWING Single-Wind Bias Data The HOW platform collections which experienced s.w. bias on Shots A, B, and D (little fallout on Shot B) are illustrated in Figs. 3, 5 and 7. Wind directions and observed bias directions are depicted for comparison and are discussed below. As the six sample values for a platform were insufficient to approximate the collection curve, the curves that are shown in Figs. 4, 6 and 8 were completed through interpolation based on the requirement of symme-trical variation about a minimum-maximum axis as derived from the air-flow studies. Curve fitting was done by trial and error methods in which the locations of the maximum and minimum collections (180° separation) were first assumed. With this assumption the locus of maximum and minimum values each form a line perpendicular to the abscissa. From a probable point on the maximum value Locus and symmetric about this locus, two diverging lines of best fit were extended through the data points until they intersected the minimum value locue. Near this intersection and the forementioned originat~- ing point the two lines were further fitted for continuity (curved portions) as consistent with platform geometry. Aimittedly there is a certain amount of arbitrariness about the curves, particularly the values of the interpolated maximum and minimum. Less arbitrary are the locations of these values which jetermine the observed bias direction. The agreement between the wind direction and observed bias directions indicate the curves are fairly representative and therefore typify &ew. collection curves. Though the wind directions would aid materially in defining the curves, i.e., the immediate establishment of the positions of minimum and maximum, they were purposely reserved for the forementioned test. Thus by proper interpolation ani limited sampling, it is possible to adequately approximate the collection curve. A notable aspect of the curves is their general resemblance to sine curves. The pertinent bias characteristics of the HOW collections are summarized in Table 1. The mean platform vaiue is defined as the average of ten values tion curve. The ground value is taken ar the average of the ground collection taken at 20° intervals between the msximum and minimum values on the collec- values which are listed in Table A.1. In the case of Shot C, a ground value could not be computed because rains anid unexpected water waves rendered the ground date unreliable. Particle density values are teken or extrapolated from other studies./28 An analysis of the particie size data from incremental collectors has been completed; however, mean particle sizes have not been assigned to the platforms.9 It must be emphasized thut these instruments were also mounted in the platform end therefore subject to bias effects. Liquid fallout particles were produced by Shot C; no size measurements were taken on the HOW island collection. .