1. Specialized mechanical structures produced by organisms provide crucial fitness advantages. The energetic cost associated with producing these structural materials and the resulting energetic trade-off with growth, however, is rarely quantified. We evaluate resource allocation to structural material production within the context of an energetic framework by combining an experimental manipulation with an energetic model.

2. Mytilid bivalves produce byssus, a network of collagen-like threads that tethers individuals to hard substrate. We hypothesized that a manipulation that induces the production of more byssal threads would result in increased energetic cost and decreased growth.

3. In month-long field experiments in spring and autumn, we severed byssal threads across a range of frequencies (never, weekly, daily), and measured shell and tissue growth. We then quantified the costs associated with the production of byssal threads using a Scope for Growth model.

4. We found that byssal thread removal increased byssal thread production and decreased growth. The cost calculated per byssal thread was similar in the spring and autumn (~1 J/thread), but energy budget calculations differed by season, and depended on thread quantity and seasonal differences in assumptions of metabolic costs.

5. This work demonstrates that the cost of producing a structural material has a substantial effect on mussel energetic state. The energetic cost of producing byssal threads was 2-8% percent of the energy budget in control groups that had low byssal thread production, and increased 6 to 11-fold (up to 47%) in mussels induced to produce threads daily.

6. We propose that characterizing the trade-off between the cost of biomaterial production and growth has implications for understanding the role of trade-offs in adaptive evolution, and improved natural resource management and conservation practices.

The effect of thread severing frequency on thread production rate and growth of Mytilus trossulus was investigated in a field setting over 1 month in autumn 2013 (mid October to mid November) and in spring 2014 (late April to late May). Mytilus trossulus, ranging approximately 2-3 cm length (~80 to 200 mg dry weight tissue), were collected from Argyle Creek on San Juan Island, WA (Lat 48.521652°N and Long 123.014061°W) and transported to Friday Harbor Laboratories (Lat. 48.525350°N, Long. 123.012521°W). The pre-existing byssal threads were severed from each mussel using scissors before the mussel was placed in a flexible mesh enclosure (10 cm x 22 cm, HDPE vexar plastic, 1 cm² mesh size) suspended from a floating dock at ~1m depth. The three treatments differed in the frequency at which the newly produced byssal threads were severed during the experiment: never, weekly or daily (or 0, 1 and 7 times per week, respectively). Fifteen mussels were distributed evenly among three replicate enclosures for each treatment.

Mussels in the “daily” treatment group were monitored for byssus production every day by counting and severing newly produced byssal threads. New byssal threads were also counted and severed for mussels in the “weekly” treatment group, but at a lower frequency of once per week. New byssal threads of the mussels in the “never” treatment group were not severed and were counted at the end of the 4-week experimental period. Thread production of each individual mussel was counted in all treatments, however in the “never” control group in spring only the total number of byssal threads produced by the group was recorded. This value was divided by sample size to obtain an average thread production for each individual in this single group.

Shell growth was calculated as the change in shell length, measured with calipers (±0.1 mm). Buoyant weight (±0.001 g) was determined at the beginning and end of the experiment. At the end of the experiment, gonad and somatic tissues were removed from the shell and dried at 60°C to a constant weight, and the dried shell weight was measured (±0.01 g). Condition index (CI) was calculated by dividing dry tissue weight (g) by shell length cubed (mm³).

In the spring experiment seven mussels either were lost from the field enclosure or died during this experiment and were not included in the analysis (ID's 11, 31, 3, 44, 54, 88, 19). In the autumn experiment two mussels died and their final length and tissue mass at 29 days was estimated from their 14 - 21 day growth rate, and the relationship between tissue weight and length cubed for the population (ID's 22, 37). In the spring experiment a single thread count was determined for the 'never' treatment population rather than for each individual (ID's 9 - 59, "thread_count").