The importance of matching the receptacle size to the size of the sample is illustrated by measurements of errors in volumes measured as a function of ball-to-receptacle diameter ratios (Figure 3A). As expected, higher ratios were generally associated with lower errors. Although individual differences in percent errors were not significant, the pooled errors for ratios >0.8 were significantly smaller than for ratios <0.8.
The importance of the MVM resolution index for measurement accuracy is illustrated by tests that used different syringe sizes to measure 0.79- and 1.0-mm diameter balls (Figure 3, B and C, respectively). Higher-resolution index values were associated with smaller mean errors and lower variability. Also, these data show that highly accurate measurements are possible even if the ball-to-receptacle diameters ratio is not particularly high (see legend of Figure 3, B and C for details). The measured differences in absolute errors were significant for the 0.79-mm balls (P =0.001) but not for the 1-mm balls. Together, these tests involving receptacle diameter ratios and the MVM resolution index confirm that both parameters influence measurement accuracy and should be maximized. Nevertheless, for any series of measurements and MVM configurations, accuracy across the expected range of sample volumes should be pretested with standard reference volumes.Overview of MVM accuracy
Figure 4 illustrates the accuracy of volume measurements that can be achieved by applying our methods across a range of standard volumes from ~1 mL down to 0.066 µL. The same low error rates that were achieved for larger volumes mainly by monitoring the meniscus level were maintained for smaller volumes by applying the reflected light method (Figure 2A), monitoring evaporation (Figure 2B), and maintaining relatively high receptacle-to-ball diameter ratios and a high resolution index (Figure 3). The standard deviations in measured volumes (Figure 4A) were all <4% of the corresponding volume, and the mean percent errors between measured and expected volumes (Figure 4B) were less than ±2% and also exhibited consistently low variability.
Validation of MVM for measuring soft tissues
Table 1 compares volumes of two brains that were successively measured by confocal microscopy and in the MVM. The estimated volumes differed by ≤10%, confirming the feasibility of the MVM for measuring volumes of very small, delicate tissue samples. For measuring volumes of freshly dissected tissue samples, the MVM was also more effective and far less time-consuming. Unfixed tissues in saline are often fairly opaque, rendering confocal microscopy impractical for all but very small specimens. The unfixed ant brains used here could only be imaged to a depth of approximately 150–250 µm (slightly over 50% of the tissue depth), even when using a high photomultiplier gain and the maximum laser intensity. Thus, to complete a series of optical sections, it was necessary to turn over and realign the brain. Using confocal microscopy to measure larger or thicker tissues would require dehydrating and clearing the tissue, therefore shrinking it. A second advantage of the MVM was that volume measurements were obtained much faster. After dissection, we estimated the total time required to measure one brain with the MVM to be <20 min versus ≥4 h using confocal microscopy and profile tracings. Traditional histology would require even more time.
Comparisons to alternative methods
An obvious alternative to volumes is to measure mass, as is common in studies of vertebrate brain-body scaling (2). Since the densities of water and many biological tissues are similar, mass and volume data are often roughly comparable. Microbalances capable of matching the MVM's accuracy of 0.06 mg ± 2% are expensive, however, and cannot be used under rugged field conditions. Another alternative (also unsuitable for remote locations) is to exploit the relationship between an object's volume and its weight, as indicated on a balance, when the object is held suspended in a fluid (16). Although this method is easily applied to relatively large volumes, significant challenges arise for smaller volumes, because it depends on the density difference between the sample and the fluid, which is typically very small (17,18).
Other precise methods of measuring small volumes exist, such as reconstruction of volumes from histological, optical (confocal), or even NMR sections (19), but such equipment is not portable. Such procedures (e.g., histology) are also subject to artifacts such as tissue shrinkage (20), which may reduce volumes by as much as 40–49% (21). The extent of shrinkage may differ according to taxonomic, developmental, and other variables, and shrinkage can even vary significantly among identically processed tissues (22).
The MVM and procedures presented here provide a rapid and effective means of measuring living organisms or freshly dissected tissues having volumes as small as 0.06 µL. The method also can be used in conjunction with histological studies to establish correction factors for changes in native volumes.