where Py is the value of the grain diameter in mm. at a given percentile 
of the grain-size distribution. The measure of sorting is one used as 
a matter of convenience in terms of the settling tube. The estimate of 
mean nominal diameter used employs the same values used in determination 
of So and, in addition, is considered relatively "efficient" (McCammon, 
1962). 
ORGANIZATION OF DATA 
The forty-one sampling stations in the vicinity of Rudee Inlet were 
numbered serially and plotted on the base map (Figure 7), using a U-V 
co-ordinate system (Krumbein, 1959) that was geographically true. The U 
co-ordinate starts with U = O at the top of the swash—backwash zone in 
the northwest corner of the map and proceeds south along the shore to 
U = 553 at the top of the swash in the southeast corner. The V co-ordinate 
extends from V = O at the northwest corner to V = 300 in the northeast 
corner. Water depth at this last point was 8 feet. Each station was 
numbered and the values for Mz and So were listed for each station. 
These data comprise an irregularly spaced set of stations; i.e., the 
sampling points do not lie on a rectangular grid. 
Data for the three transects were treated separately. Stations were 
also numbered serially and coded in true geographic (U, V) co-ordinates. 
The origin of this system was at the most inshore station of Transect A 
(Figure 5), with the U axis extending southward parallel to shore. The 
sampling points of this second system lie at the intersection points of 
an orthogonal grid. 
TREND SURFACE ANALYSIS 
A trend surface may be visualized as a smooth contour surface show- 
ing the systematic pattern of variation of a mapped variable from one 
map edge to another. This contrasts with the small-scale fluctuations 
from one point of observation to the next, superimposed on the trend as 
a seemingly non-systematic component. 
Most techniques of trend analysis are based on least squares 
fitting of polynomial surfaces to the data by a multiregression tech- 
nique. When (as is usually the case) geological observations are 
scattered irregularly over a map area, the method of analysis commonly 
includes fitting successively higher order surfaces to the map data, 
including usually the linear, quadratic, and cubic terms. These computed 
surfaces and their deviations are examined for their geological implica- 
tions. The geologist thus "takes his data apart" in various ways, 
sharpening the interpretation of the observed map data. The complete 
trend, defined by Grant (1957) as the polynomial of the best fit to the 
data, may not always be identified by these low-order surfaces. However, 
in many maps the linear and quadratic trends are strongest, with those of 
cubic and higher degree diminishing relatively in importance. 
