Let $$ (\mathbb{S}^{{d - 1}} ) = {\left\{ {x:{\left| x \right|} = 1} \right\}} $$ be a unit sphere of the d–dimensional Euclidean space ℝ d and let $$ H^{p} \equiv H^{p} (\mathbb{S}^{{d - 1}} ) $$ (0 < p ≤ 1) denote the real Hardy space on $$ \mathbb{S}^{{d - 1}} .$$ For 0 < p ≤ 1 and $$ f \in H^{p} (\mathbb{S}^{{d - 1}} ), $$ let E j (f,H p ) (j = 0, 1, ...) be the best approximation of f by spherical polynomials of degree less than or equal to j, in the space $$ H^{p} (\mathbb{S}^{{d - 1}} ). $$ Given a distribution f on $$ \mathbb{S}^{{d - 1}} , $$ its Cesaro mean of order δ > –1 is denoted by $$ \sigma ^{\delta }_{k} (f). $$ For 0 < p ≤ 1, it is known that $$ \delta (p): = \frac{{d - 1}} {p} - \frac{d} {2} $$ is the critical index for the uniform summability of $$ \sigma ^{\delta }_{k} $$ in the metric H p . In this paper, the following result is proved: Theorem Let 0<p<1 and $$ \delta = \delta (p): = \frac{{d - 1}} {p} - \frac{d} {2}. $$ Then for $$ f \in H^{p} (\mathbb{S}^{{d - 1}} ), $$ $$ {\sum\limits_{j = 1}^N {\frac{1} {j}{\left\| {\sigma ^{\delta }_{j} (f) - f} \right\|}^{p}_{{H^{p} }} \approx {\sum\limits_{j = 1}^N {\frac{1} {j}E^{p}_{j} {\left( {f,H^{p} } \right)}} }} }, $$ where A N (f)≈B N (f) means that there’s a positive constant C, independent of N and f, such that $$ C^{{ - 1}} A_{N} (f) \leqslant B_{N} (f) \leqslant C\;A_{N} (f). $$ In the case d = 2, this result was proved by Belinskii in 1996.
Read full abstract