The larvae of Dendraster excentricus orientate vertically, anterior upwards, owing to the low centre of gravity of the skeletal mass.  Forward swimming, then, leads to upward movement.  Field collections in the San Juan Islands, Washington show that 4-arm and later-stage larvae usually inhabit waters less than 6m depth.  Occupation of surface waters in this way leads to recruitment into shallow subtidal and intertidal habitats. 

Density of the developmental stages varies with skeletal mass and relative amount of lipids contained in the body. The lowest-density stages are egg through prism and highest are 4- through 6-arm larvae (see graph on Left).  Density decreases slightly at the 8-arm stage owing to accumulation of lipid droplets - energy to fuel development through metamorphosis. 

The larvae exhibit a daily vertical migration, moving a few meters down during the day and returning to surface waters during the night.  The function of this behaviour is unclear, but may relate to low tolerance of the larvae to UV-B irradiance.  While survival is good in ambient laboratory light and in normal UV-A (includes visible light), greater-than-natural intensity UV-B treatments (include UV-A and visible) over 3d are lethal (see graph on Right). The authors note that UV-B at natural intensities causes the larvae to stop swimming and descend, and they suggest that vertical migration may be a way that the larvae maintain their position in surface waters, but at depths below harmful levels of UV-B irradiation.  The authors remark that similar outdoor tests in full sunlight would be useful future research. Pennington & Emlet 1986 J Exp Mar Biol Ecol 104: 69.

NOTE neutral density is 1.0, so the later larval stages are expending much energy in swimming to maintain constant depth

NOTE  4-arm larvae are kept in shallow (<3cm deep) glass bowls and subjected to different irradiance treatments for 8h per day over an 8-d period

The tiny sizes of sand-dollar larvae and of invertebrate larvae, in general, means that viscosity¹  of water may influence their speed of movement.  Temperature affects viscosity through physical means and also speed of metabolic reactions.  Temperature can influence a swimming echinopluteus larva both physically and physiologically.  A temperature decrease, for example, will lower swimming speed, but how much of the decrease is due to increased viscosity, and how much to decreased metabolism? 

Studies on this interesting subject at Friday Harbor Laboratories, Washington using 6- to 8-arm Dendraster excentricus larvae (12-19 days from hatching at 20^oC ) and seawater with altered viscosities²  show that physical effects of viscosity can comprise a large component of the effect of temperature on activity of small organisms.  For example, a decrease in water temperature of 10^oC will reduce swimming speed of echinoplutei by 40% and water moved³  by the larvae by 35% (this is shown by the difference in the Left and Right sets of bars in the histogram).  However, 40% of the decrease in swimming speed and 55% of the decrease in water moved are accounted for by increases in viscosity alone. This is shown by the middle set of bars, which show swimming speed at 22^oC but with viscosity adjusted to that at 12^oC (the difference, then, gives the effect of viscosity alone). The authors note that these effects of viscosity are biologically relevant.  They caution that if viscosity effects are not corrected for in experiments they may lead to greatly overestimated effects of temperature on metabolic processes.  For example, calculation of Q_10⁴  for water moved at the 2 temperatures gives a value of 1.5, but if viscosity is held constant, the value is 1.2.  Thus, without including a correction for viscosity change with a decrease of 10^oC , the effect of temperature on water moved by the sand-dollar larvae will be overestimated by 23%.  Podolsky & Emlet 1993 J Exp Biol 176: 207.

NOTE¹  resistance of a liquid, in this case, to flow.  Viscosity results from the internal friction of the material’s molecules.  As temperature decreases, the viscosity of most materials increases.  Seawater in the tropics at 30^oC, for example, has less than half the viscosity of seawater in the Arctic at 0^oC.  Podolsky & Emlet 1993 J Exp Biol 176: 207.

NOTE²  viscosities are increased with additions of metabolically neutral polyvinyl pyrrolidone (PVP: Mr 360000: Sigma Chemical Co.)

NOTE³  note that changes in temperature and viscosity will affect swimming speed through changes in the amount of water moved by the cilia per unit time.  The authors could have measured either factor to test their hypothesis, but chose to do both.  To measure water moved by ciliary propulsion the larvae are held in place by a suction pipette over one of the arms and analysed using high-speed video techniques with plastic microspheres in the water (see drawing on Right). Swimming speeds are measured after 2-4h acclimation to the different temperatures with the larvae moving vertically (their natural tendency) over a 9mm distance

NOTE⁴  a coefficient that gives the relative change in a rate over a specified 10^oC change in temperature

As seen from these drawings of 2- to 8-arm larval stages in Dendraster excentricus, most of the growth after the 4-arm stage occurs in the arms.  In fact, a study at Friday Harbor Laboratories, Washington shows that growth of the 2 pairs of largest arms, the preoral and posterodorsal, is allometric, as exhibited by the scaling relationships of ciliated band lengths to various body dimensions (e.g., slope of ciliated band length to larval length = 2.1).  Overall, this change in shape during growth accounts for 83% of the increase in feeding capability from the 4-arm to 8-arm stages. 

Accompanying this allometric growth is a proportional increase in metabolic activity (see graph on Right). The author sums up by stating that an important functional consequence of the change in larval form is the maintenance of feeding capability relative to larval body size and energy demand.  McEdward 1984 J Exp Mar Biol Ecol 82: 259.

NOTE  lit. “different measurement” G., as opposed to isometric.  In the example above, if the relationship were isometric the slope would not differ significantly from a value of 1 (Y = aXb, where b is the slope of the regression line); that is, one unit increase in larval length would be accompanied by one unit increase in ciliated band length.  However, the data show that ciliated band length scales with the square of larval length (slope b = 2.1), which is allometric