The free swimming larval form, or “tadpole”, of the tunicate B. viocella is released from its colony and swims for 1 to 3 days before attaching to a solid surface where it metamorphoses to the adult form as a filter feeder in approximately 24 hours. The larvae is large and complex compared to most other tunicates, and a heartbeat can be seen. The circulatory system is highly compressed at this stage and is further obscured by the tunic gel enclosing the tadpole. However, the approximately 30 ellipsoidal ampullae are visible as a parallel bundle around the surface of the anterior. The ends of these ampullae will be the major regions of attachment to the solid substrate, after which the ampullae rapidly spread outward radially on the surface.
Even at the time of attachment several small buds can be seen on the surface of the tadpole. Each bud is developing into a new individual, or zooid, to augment the parental zooid in the growing colony. The time chosen in this report to document pulsations of the ampullae and other parts of the colony is a compromise between delaying long enough for the ampullae to be clearly visible as they spread out from under the zooid body, but before the zooid, buds, and the associated circulatory network grow and become even more complex. The exact number of ampullae, the pattern they form on the surface, the details of the connecting vessels, the size and number of zooid buds, and the periods of contractions vary from colony to colony. The first zooid of the colony is sometimes called the oozooid, to emphasize that it has developed directly from the egg. However, only young colonies, which contain only a oozooid are considered here and thus the distinction will be dropped.
General conclusions made in this report have been distilled form observation of more than 100 colonies over two years. However, the majority of actual measurements presented in the Figures come from one colony that had a combination of characteristics that facilitated observation and quantitative measurement. The focus on one colony reveals how all its parts function together.
A typical tunicate colony
The colony seen in Fig. 1 was initiated 4 days earlier by a 1 day old tadpole. This grey level image was derived from one frame of a color video sequence of 2048 seconds. The area of ampullae number 1 was obtained by a computer controlled scan documented in the insert and described in detail in in the Methods section. The solid white lines in the insert are the ends of chords that define the area.
The vessel network connecting ampullae and zooid
The vessel network that links the ampullae and zooid can only be partially discerned in Fig. 1. In order to obtain a more detailed and functional description the contrast was greatly increased using software and blood cell movement was visualized by repeated examination of accelerated (time-lapse) segments of the video.
The resulting map is presented in Fig. 2. as a net of red lines that for the most part overlay the sometimes faint grey lines seen in Fig 1. There are also small ampullae that may grow into larger ones, but their behavior is not described in this report.
The basic topology of the net is a set of radial vessels from ampullae to a vessel ring which appears here to be connected to the zooid body at only 5 locations, but here the central segment of the network is obscured by blood movement in the zooid body. Previous observations indicated that typically these vessels connect to circulation in the zooid or associated buds at 2-4 points. Note that a few ampullae are almost directly connected to each other, e.g. 20 and 21, while others are far more isolated, e.g. 15 and 16.
Contractions of one ampullae
Contractions of ampullae diameters along the tip-base axis are generally in phase, which was the case for all the ampullae that could be reliably scanned in the colony seen in Figure 1. However, for some ampullae in some colonies, contraction of some segments, a few percent of the total, can be significantly out of phase. Thus the mechanism of contraction does not absolutely ensure synchrony.
As seen in Figs. 1 and 2 the two dimensional profiles of most ampullae appear to approximate an ellipse with a length about 2-4 times longer than the width. Observation of accelerated videos suggested that width changes more than length during the rhythmic pulsations and measurement confirms this conclusion.
Actual measurements of changes in width and length of amp 1 over a period of 1024 seconds are seen in Fig. 4. Though the ampullae’s width is only about 1/3 its length, the width changes by about twice the absolute number of pixels. Changes in the widths of ampullae 1, 19, and 25 were found to have a range of 37, 27, and 23 percent, while the lengths had a range of 7.5, 5.3, and 8.7 percent respectively. Thus rather than being similar to elastic balloons that expand and contract symmetrically, ampullae are more similar to elastic cylinders that expand and contract mostly in radius. One consequence of this asymmetry is that the measured changes in area of the profiles are fair approximations to changes in the surface area of the ampullae.
Fig. 5 is a plot of the contraction profile of ampullae 1 for 2048 seconds and a sine wave picked by eye to match the regular parts of the ampullae curve as best as possible. This sin wave has 25 cycles in 2048 seconds, thus a period of 82 seconds. The average period of ampullae contractions in 12 other colonies was found to vary from 62 to 143 seconds, with a mean of 104 seconds, thus the period of the colony seen in Fig. 1 is fairly typical. While the area profile of ampullae 1 appears similar to the sine wave in many segments (that is how this sine wave was chosen) in the ampullae profile differences between peak times are not constant, peak shapes change, and peak extremes are not constant.
An intuitive way to characterize an oscillating curve is to measure the times between the maxima, the peak to peak times. The average peak to peak time for ampullae 1 is 76 seconds, quite close to the constant 82 second peak to peak time of the sine wave, but the standard deviation of the ampullae times is 22 percent, because most ampullae peaks often come slightly before or after the peaks of the sine wave.
A more complete comparison between the contractions and sine waves can be obtained by computing the Fourier power spectrum (FPS), seen in Fig. 6, which is sensitive to peak times, heights, and shapes. The major peak for ampullae 1, at a frequency of 26 per 2048 seconds (period of 79 seconds) is only 15.9 percent of its total power, with the remaining distributed over many higher and lower frequencies. As expected (or defined) the power spectrum of the sine wave with that frequency has but a single peak containing 100 percent of the power at this frequency.
Contractions of a group of ampullae
To see variation and search for patterns of similarity between contractions of many ampullae with respect to both position and time we need to be able to see a lot of data in one graphic, but not with great precision. Coding ampullae areas as colors makes this possible. The areas of 15 adjacent ampullae, more than half the colony perimeter, over a time span of 2048 seconds are seen in Fig. 7 as 15 parallel vertical time lanes, with yellow representing large, green average, and blue small areas.
Ampullae 14 and 15 have similar patterns. Ampullae 20 and 21are seen to share many dramatic blue bands, for example at 200, 1050, 1200, 1570, 1800 and 1950 seconds. This is not surprising since each of these pairs are neighbors connected by a short vessel segments as seen in Fig. 2. Ampullae 16, 17, and 18 have similar large peaks for the first 800 seconds, but share some, but not all, blue bands after that. Lanes 9 and 10 share many bands. In general neighboring ampullae tend to be similar, and distant ones different.
However, the lack of a simple or even symmetric relation between contractions of most ampullae is consistent with the complex and mostly non-symmetric geometry of the vessel net connecting the ampullae with each other and the zooid. It is interesting that adjacent segments of distinct peak patterns do not seem to “diffuse” laterally to neighboring ampullae, they just come and go in their own neighborhood.
The periodicity of each of the 15 ampullae can be assessed in Fig. 8, a color map of their Fourier power spectra. The total power in the spectra displayed in this map have not been normalized for each ampullae, thus ampullae that have more extreme contractions have higher total values over all the frequencies. This is illustrated by more red in lane 21 compared to lane 9. The considerable differences between ampullae are clear.
Correlations between pairs of ampullae
Only an intuitive feeling for correlations of contractions between the 15 ampullae can be obtained by examination of Figs. 7 and 8. In contrast the actual values of all the possible 105 pair wise correlations of contractions are represented as colors in Fig. 9.
Three red 2x2 squares along the diameter of Fig. 9 indicate the high correlations between the ampulla 10-11 pair, the 14-15 pair, and the 20-21 pair. In contrast, the blue rectangles toward the upper right and lower left corners represent the high negative correlation between the 20-21 pair with ampullae 8 through 12. Ampullae that are neighbors are more likely to be correlated than pairs that are distant, but the detailed pattern is complex and any other generalization is not obvious.
Rolling time correlations between ampullae
A rolling time or time windowed correlation between two sequences is a series of correlations between two small subsets of the sequences that moves along in time. It thus reveals transient time dependent correlations that are averaged out when the entire series are compared.
It is seen in Fig. 10 that the two ampullae are highly correlated in a 60 second window for the first 900 seconds, then become much less for 100 seconds, then return to high correlation for another 500 seconds. The remainder of the time they abruptly go into and out of correlation several times. A sequence of high and low correlation periods is characteristic of two oscillators that are partially entrained.
Phase portraits of two ampullae
The phase portrait, showing the trajectory of an objects motion, or behavior, in phase space, is a common tool used to understand dynamical systems. It can reveal subtle aspects of a system and is not linked to a specific function, as Fourier analysis is. The top panel of Fig. 11 shows the plot of area versus time for ampullae 9, with the brown segment indicating the segment to be plotted as a phase portrait in the lower panel. The curve displayed there, starting with the green and ending with the red segment, is the ampullae area versus the time derivative of that area. If the time axis extends upward from the paper, this curve is a helical trajectory collapsed on the page.
The trajectory for a pure sine wave would be a series of superimposed ellipses, one for each complete cycle of the wave. If the area of the ampullae was suddenly altered by an external force, but the system was stable, the trajectory would gradually return to the original elliptical curve, which is thus called an attractor. In contrast, the trajectory for ampullae 9 is a series of loops of increasing size reflecting the increasing size of the cycles in the top panel. The trajectory is flat on the left (most negative values of area) because the derivative changes a great deal over a small range of area, the actual data curves are sharper on the bottom. A model for ampullae 9 would be an oscillator with an intrinsic frequency being influenced by modest, gradually increasing external forces causing excursions from the intrinsic ellipse.
The trajectory for ampullae 21 over the same time period, seen in the bottom panel of Fig. 12, is quite different, a reflection the different pattern of the raw date seen in the upper panel. Three of the trajectory loops have small subloops or extrusions. These correspond to small inflections or peaks in the data. Thus strong but transient forces are modulating the behavior of this ampullae.
Vessels connecting ampullae also contract.
Observation of time lapse versions of videos of colonies revealed that the widths of all vessels connecting the ampullae and zooid body were contracting along with ampullae contractions. A scan of a vessel near an ampullae, seen in Fig. 14, confirmed and measured these contractions. In fact the percent change in the vessel width is perhaps slightly larger than the change in ampullae area for this specific region.
The sum effect of all ampullae in the colony
Since contractions in most ampullae are not in complete synchrony, the sum effect of these contractions, pumping blood in and out of the zooid body, would be expected to be less than maximal because some flow will just be from one ampullae to another. It was possible to scan 19 of the 27, ampullae, with good sampling around the periphery of the colony. The sum of all 19 was a curve very similar to that seen previously for individual ampullae, with a range of 19.1 percent from the mean. However, the mean of the range of individual ampullae was 30.5 percent. Thus lack of complete synchronization has reduced net pumping to 63 percent of the possible maximum.
The zooid body and its peristaltic heart
In preceding sections it has been shown that contractions of the ampullae and associated vessels are sufficiently synchronous to produce a net flow in and out of the zooid body with a period of about 80 seconds for this colony. This flow must cause parts of the zooid to swell and contract with the same period but opposite phase. However, the zooid is not a solid mass of tissue, rather it can be approximated as a small posterior region containing viscera plus two large concentric cylinders enclosing sea water. The outer cylinder is the body wall and the inner cylinder the brachial basket that supports the mucus feeding net. Thus, while some portion of the body must expand and contract, all that is clearly visible in the videos is the outer profile of the zooid.
However, measurement of the area of this outer profile shows that in fact it does contract rhythmically with the same frequency but mostly 180 degrees out of phase with ampullae contractions, as seen in Fig 14.
A peristaltic heart in the posterior region of the zooid body, with a beat rate approximately 100 times faster than contractions of the ampullae, circulates blood throughout the zooid. This heart periodically changes direction, which can be identified, in favorable cases, by viewing videos of this colony. In Fig. 15 the direction of blood flow generated by the peristaltic heart is represented by the rectangular curve alternating between 100 percent in the anterior direction along the endostyle, and 100 percent to the posterior. Contractions of the ampullae are superimposed for comparison. Starting at the left end of the graph it is possible to associate a minimum of the ampullae curve with a heart flow transition across most of the graph. Thus, in this colony, flow due to the peristaltic heart appears to be entrained with flow due to ampullae contractions, but with twice the period.
Contractions of a detached pair of ampullae.
In 1899 Frank W. Bancroft published observations on contractions of ampullae of colonial tunicates in an report[1] that is quite accessible today. He also described contractions in groups of ampullae that had been surgically separated from colonies.
Fig. 16 describes the contractions of the most simple group, a pair. These detached ampullae were connected to each other by a short vessel segment, similar to that seen between ampullae 20 and 21 in Fig. 2, however they came from a different colony. The uniformity and symmetry in these pair of contractions is apparent to the eye. The average of the peak to peak times was 127 seconds, with a standard deviation of 9.4 percent. The curves are almost indistinguishable from sine waves, with 94 percent of the total power spectrum in the main frequency.
Contractions of single detached ampullae
Even single detached ampullae contract with periods similar to those observed when they were attached to a colony. While the total volume can’t change in one isolated ampullae because blood is incompressible, the ampullae can and does oscillate as can be seen by blood cell movement and by changes in shape as documented in Fig 17.
In all ampullae described in this report changes in shape or size were correlated with movement of blood cells. If ampullae are part of a colony blood cells move in or out of the ampullae, in a detached pair blood cells move from one ampullae to the other; in detached single ampullae blood cells move from one end to the other. Blood cell movement is usually the only way to see changes in shape or size when viewing in real time with a microscope because changes in shape and size are too slow for the human to perceive directly.
The image of a single detached ampullae, the fifth in a series of 6, is seen in the upper left panel of Fig. 17. A segment of this and a another ampullae in the series were scanned and the results are presented in the middle panel. Note that there is nothing fundamental about this segment, it was chosen merely because it changed the most, other ampullae change shapes in different patterns.
The time course of contractions of the series of 6 were followed for approximately 1000 seconds. In contrast to previous observations, the results of these scans are not areas of ampullae, but just surrogates to follow oscillation. Thus, only the times of peaks have obvious meaning. The mean periods and standard deviations of peak-peak times are seen in Table 1 and the curves for ampullae 1 and 5 are presented in the lower panel of Fig. 17.
Table 1. The means and standard deviations of peak to peak times for 6 single detached ampullae.
Ampullae
|
mean (sec)
|
stan dev (pc)
|
1
|
68.2
|
6.4
|
2
|
67.3
|
5.0
|
3
|
65.9
|
6.3
|
4
|
68.6
|
7.9
|
5
|
68.0
|
4.0
|
6
|
68.4
|
3.7
|
The first and fifth ampullae seem so similar as to invite a comparison. Thus, the first peak of each was aligned and the two plotted together to make the bottom panel of Fig. 17. While the peak heights of each of the two ampullae change with time and pattern, the peak times match up to an amazing extent, with the 14th peak of both matching within 3 seconds out of the total of almost 900.