TABLE 1 .

Symmetry determination of NEC rings^a

Radius (nm)	Orders significant by t test	Order (SRP value)
5.04		20 (5.60E–11)
		2 (3.26E–11)
		32 (8.93E–16)
6.16		36 (7.29E–20)
		34 (6.55E–21)
		2 (7.91E–22)
7.28	7	7 (1.02E+12.00)
		4 (2.41E–19)
		2 (1.05E–25)
8.40	7	7 (3.06E+12.00)
		4 (4.34E–14)
		6 (1.09E–21)
9.52	6	6 (25.03)
		7 (5.64E–8)
		4 (2.51E–17)
10.64	6	6 (6.25E+16.00)
		4 (7.85E–25)
		28 (1.00E–25)

^aIn ROTASTAT (29), significant orders of rotational symmetry are detected in two ways. In one approach, the spectral ratio product (SRP) is calculated whereby the ratio of the amplitude of a given harmonic at a given radius and the amplitude of the same harmonic for a similarly normalized piece of background are calculated for each ring in the data set (n = 188 in this case). These ratios are combined multiplicatively to give the SPR. If the ratios are, on average, even slightly less than unity, the SPR decays rapidly to very low values. With valid symmetries, the SPR remains above unity and may diverge to high values. The other approach is based on the Student’s t test. All orders of symmetry deemed statistically significant at a specified level (we used the conservative cutoff of 10⁻⁶) are output. The two approaches gave consistent results. Significant orders of symmetry in-ring are bolded. Number of particles, 188; number of backgrounds, 188; successive rings are at radii 2 px apart. Significance level = 1.0E⁻⁶. Orders with the three highest SRP values are given.