1: \begin{abstract}

2:  Fractal Interpolation Functions (FIFs) developed through

3: Iterated Function Systems (IFSs) offer more versatility than the

4: classical interpolants. However, the application of FIFs in the

5: domain of shape preserving interpolation is not fully addressed so

6: far. Among various interpolation techniques that are available in

7: the classical numerical analysis, the rational interpolation

8: schemes are well suited for the shape preservation and shape

9: modification problems. Consequently, we introduce a new class of

10: rational cubic spline FIFs that involve tension parameters as a

11: common platform for the shape preserving interpolation and the

12: fractal interpolation to work together. Suitable conditions on the

13: parameters are developed so that the rational fractal interpolant

14: retain the monotonicity and convexity properties inherent in the

15: given data. With some suitable hypotheses on the original data

16: generating function, the convergence analysis of the rational

17: cubic spline FIF is carried out. Due to the presence of  scaling

18: factors in the rational cubic spline fractal interpolant, our

19: approach generalizes the classical results on the shape preserving

20: rational interpolation by Delbourgo and Gregory [SIAM J. Sci.

21: Stat. Comput., 6 (1985), pp. 967-976]. Furthermore, for preserving

22: shape of a data set wherein the variables representing the

23: derivatives have varying irregularity, the present schemes

24: outperform their classical counterparts. Several examples are

25: supplied to support the practical value of the method.\\

26: {\bf This paper was submitted to SIAM Journal of Numerical Analysis on April 14, 2013 after initial submission to Constructive Approximation on July 18, 2012.}

27: \end{abstract}

28: