Jmp 7

In order to characterize the variation in floral morphology among natural populations of S. rostratum, we collected floral morphology data from six populations across a latitudinal gradient in Mexico during October and November of 2010 (Table 1). In each population, we measured between two and four flowers from 16 to 30 individuals (Table 1). For each flower, we measured the following ten traits with digital calipers: corolla length (1) and width (2); the length of the anther and the width of the anther at its widest point, for both the feeding (3, 4) and pollinating anther (5, 6); the length of the style (7); the distances between: the stigma and the pollinating anther (8), the stigma and the nearest feeding anther (9) and the pollinating anther and the nearest feeding anther (10; Figure 1). We analyzed these floral measurements using principal component analysis (PCA) of a correlation matrix. Differences among populations in the first two principal components were analyzed using an analysis of variances (ANOVA) of the principal component scores, and a Tukey post hoc test. For this analysis, we used JMP 7.0.2 (SAS Institute Inc. 2007) and plotted the results with Sigmaplot 13 (Systat Software Inc. 2015).

Statistical analyses were performed separately for each fish species, with a focus on intraspecific differences in blood chemistry values between seasons. Blood-based metrics were normally distributed and compared using a two-way analysis of variance (ANOVA) with treatment and season as main effects, and treatment × season as an interaction term. Tukey's post hoc test was performed when at least one main effect or the interaction term was deemed significantly different. Data analysis was completed using JMP 7.0.2 (SAS Institute, Cary, NC, USA) with α = 0.05.

Group comparisons were performed using a one-way analysis of variance (ANOVA) and Student's t-test. Significant levels were set at P < 0.05. Group data are expressed as means ± S.D. Correlation of infection levels and mucus antibody in individual fish was performed using Kendall's τ. All calculations were performed using the statistical software JMP-7.0.2 (SAS Institute, 2007; Cary, NC).

The patency rate was estimated using the Kaplan–Meier technique. The difference in the patency rate between the two groups was examined using a log-rank test. A univariate analysis of categorical variables was made with a Chi squared test and a multivariate analysis of AVF survival was examined by the Cox proportional-hazards model. Data were expressed as mean ± standard deviation (M ± SD). p < 0.05 was considered significant. All statistical analyses were performed using the Windows version of JMP 7.0.2 software (SAS Institute Inc., Cary, NC, USA).