The actin-, myosin-, and calmodulin-binding protein caldesmon (CaD) is expressed in two splice isoforms: h-CaD, which is an integral part of the actomyosin domain of smooth muscle cells, and l-CaD, which is widely expressed and is involved in many cellular functions. Despite extensive research for many years, CaD's in vivo function has remained elusive. To explore the role of CaD in smooth muscle contraction in vivo, we generated a mutant allele that ablates both isoforms. Heterozygous animals were viable and had a normal life span, but homozygous mutants died perinatally, likely because of a persistent umbilical hernia. The herniation was associated with hypoplastic and dysmorphic abdominal wall muscles. We assessed mechanical parameters in isometrically mounted longitudinal strips of E18.5 urinary bladders and in ring preparations from abdominal aorta using wire myography. Ca2+ sensitivity was higher and relaxation rate was slower in Cald1−/− compared with Cald1+/+ skinned bladder strips. However, we observed no change in the content and phosphorylation of regulatory proteins of the contractile apparatus and myosin isoforms known to affect these contractile parameters. Intact fibers showed no difference in actin and myosin content, regardless of genotype, although KCl-induced force tended to be lower in homozygous and higher in heterozygous mutants than in WTs. Conversely, in skinned fibers, myosin content and maximal force were significantly lower in Cald1−/− than in WTs. In KO abdominal aortas, resting and U46619 elicited force were lower than in WTs. Our results are consistent with the notion that CaD impacts smooth muscle function dually by (1) acting as a molecular brake on contraction and (2) maintaining the structural integrity of the contractile machinery. Most importantly, CaD is essential for resolution of the physiological umbilical hernia and ventral body wall closure.