RRWPC Comments on 2019 NCRQCB TMDL Action Plan, 6/24/19

Comments on Staff Report Revisions (May, 2019)Action Plan for RR Watershed: Pathogen TMDL

As much as feasible, we have tried to limit comments to revisions only.

Hours before I completed these comments, I received the comments written by Rich Holmer.  As an expert in the field of septic system management and public health, he expressed many of the concerns I shared (and more) far better than I could have done. I fully support his comments and admire his support of the needs of the community and individual property owners.   I submit my comments as well in the hopes it can add to his fine work.

In looking at today’s Press Democrat (6-23-19) and this week’s Sonoma West Times and News, it is clear that issues and concerns discussed over the last several years about this project, are coming to a head.  The Action Plan makes no adequate attempt to address the questionable need for perhaps unduly expensive professional inspections, lack of certainty regarding direct funding assistance for low income people (other than a list of potential funders), substitution of bacteroides to demonstrate  water quality impairment in the lower main river rather than limited E. coli (for fresh waters) and enterococcis (for saline waters), equating bacterial pollution in the tributaries (lower river: Green Valley, Dutch Bill, etc.) with supposed pollution in the main stem Russian River, etc.

Descriptions of the lower river area in the document state that it’s a densely populated rural area.  This may not give a complete impression. There are numerous narrow canyons that have houses lining them on either side of the road with no houses behind them.  At the end, the road may circle and come back again, but there’s only one line of houses on one or both sides of the road on either side of the canyon.  (In some areas parks and empty lots line the street.) As you go further back into the canyon, many of the parcels are much larger.  Many of these houses are not occupied year round and  many are rather small (I’m guessing 1000 sq. ft. to 1500 sq. ft. on 5,000 sq. ft. lots) The impression of density is because the lots closer to the river are small and often,  back up into the hills.  It’s nothing like a city where you have block upon block of houses, including many multiple units, apartments, duplexes, etc. giving a much higher density and covering a much greater area.  And surrounding all of this is a forest of trees, almost everywhere you look.  (There are also many more open areas and houses lining the river front have beautiful views.

Some Historical Background:

As a result of the Biological Opinion and the effort to permanently lower Russian River flows, we have been aware of water quality studies in the lower river for the last nine years demonstrating E. coli exceedances of only 7% total, in over a dozen lower river locations, with most of those occurring after Labor Day and during changing river conditions (summer dams come down and river mouth closes frequently), information that has been ignored.

Furthermore, there has been little attempt to discriminate between pathogens from septics and other sources, such as recreational activities and homeless (and illegal) encampments, even though a special study documented that heavy beach use corresponded to increases and decreases of pathogens in the river at Healdsburg and Monte Rio Beaches.

Furthermore the Regional Board, the press, the Water Agency, and other institutions have appeared to ignore information provided in a formal complaint that RRWPC submitted concerning three Sanitary Sewer Overflows last February in the Russian River County Sanitation District (RRCSD) Collection System, (concerning combined spills of about a million gallons of pathogen loaded raw sewage into the Russian River), the illegal discharge of treated wastewater between May 20th and 27th, and an electrical outage at the Treatment Plant creating the need to pump sewage at Vacation Beach with a truck because no working backup power was initially available.  In spite of extensive efforts to get the word out by myself and Esa Day (neighbor of the Vacation Beach Pump Station), there were no articles in any printed newspapers and public notice was inadequate or lacking.  This is important because one of the three potential solutions to address this supposed TMDL study, is to hook Monte Rio and Villa Grande to the RRCSD.

Since most bacterial pollution occurs during big storms in winter, we may only be trading one major source of pollution for another and at great expense to home owners, many of whom cannot afford improvements as in the Action Plan.  The Staff Report expresses HOPE that this project will improve water quality, not certainty.  RRWPC supports affordable inspections in general and required replacement and repair of septic systems when merited.  We believe that most citizens want to be responsible and protect water quality and the environment, but not if it costs them their home to do it.  But we also believe that everyone is ignoring the known condition of periodic floods in the lower river, and that any remedy is probably bound to fail under the conditions we had last February.  When the central sewer system fails, and raw sewage enters the river, it pollutes with wastes from properties outside of the floodplain, as well as with those within it.


Pg. 1-4  This is the third round of comments on this project.  We had been promised that the responses would become available, but the Regional Board’s response to previous comments was not available to commenters as of the time of this deadline (June 24, 2019) even though completed.  This makes it difficult to see what changes came about as a result of those comments.  While I will attempt to only address altered text, that becomes difficult with so many issues intertwined.

Pg. 1-5: As you mention here, the 2006 Bacteria listing between Fife Creek and Dutch Bill Creek was based on Fecal Coliform.   The new State standard calling for the use of E. coli limits to determine impairment is met similarly in the 2019 Action Plan by combining fecal coliform with incidences of beach closures.  The current Action Plan calls for using beach closures with fecal coliform (where E. coli data not available) from 2012 on.  As I recall, but for one time in 2017 in the lower river (Fife Creek to Dutch Bill), the only closure was in 2017 at the Monte Rio Beach after an extremely over populated July 4th holiday resulted in high Bacteria readings that quickly went down when everyone left.  To use that data as evidence to imply that septic systems are involved, seems inappropriate.

A study published in November, 2013, on Beach Recreation Impact Report, reported the following findings (p. 7) based on samples taken at Healdsburg Beach and Monte Rio Beach over two holiday weekends.

  • USEPA (2012) Beach Action Values were exceeded for both E. coli and Enterococcus bacteria on July 2nd at Monte Rio Beach and July 4th at Veteran’s Memorial Beach in Healdsburg.
  • Significantly higher E. coli and Enterococcus bacteria concentrations were observed on July 2nd at Monte Rio Beach and on July 4th at Veteran’s Memorial Beach compared to other days during the sampling period.
  • Significantly higher Bacteroides bacteria concentrations were observed on July 3rd at Monte Rio Beach and on July 5th and July 6th at Veteran’s Memorial Beach compared to other days during the sampling period.
  • Significantly higher E. coli and Enterococcus bacteria concentrations were followed the next day by significantly higher Bacteroides bacteria concentrations at both beaches.
  • coli bacteria concentrations showed a relatively strong association with Enterococcus bacteria concentrations at both beaches. Human host Bacteroides bacteria concentrations showed a relatively strong association with All Bacteroides bacteria concentrations at both beaches.
  • Enterococcus bacteria concentrations measured at Monte Rio Beach showed a relatively strong association with Russian River estuary water surface elevation, with higher concentrations observed during higher water surface elevations.

It would seem that these findings demonstrate a high level of bacteria concentration from people recreating at these beaches.  They are the two most popular and most utilized beaches on the river.  If E. coli and enterococcus are much better indicators for pathogens, why do Public Health Departments (State and local) still rely on fecal coliform as a measure of pollution causing adverse health impacts?  What is their rationale?  Why can we not recall seeing Bacteroides as a measure of pathogenic concern by health agencies?  Have other areas in the State relied on Bacteroides as a pathogen indicator?

Middle paragraph on p.1-5 indicates that the data reanalysis relied on for this document is the same that will appear in the 2019 303(d) list analysis later this year.  Does that mean that if this document is approved with that analysis, it will affect the decision process for the 303(d) determination later this year?  Is the “tail wagging the dog” here?

  1. 1-8: It would have been helpful if you had the HUC-12 and HSA map were available side by side so one could compare the change in boundaries.  It’s hard for the lay person to visualize the difference.


This section starts on page 2-4. On page 2-6 they talk about the Potter Valley Project and Dam operation.  Nowhere in this section are the Biological Opinion (BO) and Temporary Urgency Change Orders (TUCOs) mentioned nor the Fish Flow Project, proposed to permanently lower flows in the Russian River.  The TUCOs have lowered flows between May 15 and Oct 15 almost every summer since the BO was released by National Marine Fisheries Service in September, 2008 and some before.  These have had a major impact on lower river flows and probably on bacterial pollution, which I don’t think this document addresses. What is the impact of higher and lower flows on bacterial concentrations?  (It is mentioned in the Beach report that more bacteria were at Monte Rio Beach when the river came up, which it does regularly at high tide and in fall when mouth closes more frequently.  The Estuary is a sink for a great deal of bacteria and it would be no surprise if some of it backed.  RRWPC has taken pictures from the Monte Rio Bridge that demonstrate within less than a day, the river goes down considerably when the mouth suddenly opens.  Map on page 2-8 is impossible to read. (maybe I just have a bad copy.)


Page 3-8 states (top of page): “The TMDL no longer relies upon fecal coliform bacteria data for assessment of impairment or pollution.”  But the new E. coli standard has not been approved as yet for specific 303(d) listings, so is it legitimate to say that?  This document promises to make the change, as noted earlier, but it has not gone through the approval process as yet.

And then, if E. coli is the standard, what if exceedances are mostly found during and after big storms and heavy rainfall?  If you rely on bacteroides to verify pollution and its probable source, and conclude that the presence of bacteria indicates the possible/probable presence of pathogens, how does that verify that E. coli in fresh water is polluting the river when people are recreating?  (In fact, the E. coli standard is based on causing 32 illnesses per 1000 recreators.  If you never have 1000 people exposed in the dead of winter during a big storm, does that make the measure invalid?  If you have 32 people swimming in the river during the dead of winter, would all 32 have to get sick from exposure to E.coli for it to be a valid measure?  In our view, if they are in the water in the middle of a big storm, something other than bacteria is going to harm them.  (I understand that esoteric calculations go into making these calculations on rates of illness.  Is there any layperson friendly explanation that you can provide in its stead?

In a practical sense, it sounds a little silly to insist that you have to protect REC-1 swimmers from bacterial pollution in the dead of winter and in the middle of a storms. Is it possible that information is being presented in a way to come to a desired conclusion, that the river is polluted in summer from constituents more readily present in winter (i.e.,bacteria)?  Are samples taken in the dead of winter on the off chance that we need to protect swimmers from freezing cold, fast moving, dangerous water?  Don’t we want to discourage such activities?

Another concern involves the convergence of multiple sources of pollution in relatively close proximity, such as the lower river homeless situation, the huge numbers of summer visitors at lower river beaches (see above analysis), and the existence of moderately dense housing on the banks of the river utilizing central sewer or septic, depending on their location.  How do you allocate appropriate responsibility for each group since they are all human pollution sources?

Another situation that I have not seen addressed anywhere is as follows:  The Beach Study noted that bacterial levels went up at Monte Rio as follows: “Enterococcus bacteria concentrations measured at Monte Rio Beach showed a relatively strong association with Russian River estuary water surface elevation, with higher concentrations observed during higher water surface elevations.”  The Fish Flow DEIR stated that when mouth was closed, effects could be seen as far upstream as Vacation Beach.  How are we to know whether bacterial monitoring was conducted at Monte Rio during high or low tide and/or open or closed mouth?  So another question arises: Are there other natural circumstances that could affect the amount of bacteria in the waterway that does not originate from the area in which it is found?  If so, how does that affect the validity of the winter data used to justify the need?  (also see above)

Chapter 4: Evidence of Pollution

On page 4-3 the Staff Report acknowledges that most of the recreation occurs during the summer but that REC-1 is a beneficial use throughout the year.  On what real life basis do you draw this conclusion?  Have you found people swimming in the river when the temperature is 40 degrees or when heavy rain is pounding the river’s surface?  The Report notes that monitoring takes place at all times of the year.  Should monitoring results consider, as part of the analysis, the date on which samples are taken and weather conditions?

This Report keeps mentioning the fact that one of the reviewers of the proposed TMDL thought that enterococci could be considered a pollution source in fresh water, if it is backed up with other evidence such as beach postings.  Does anyone at the State concur with this view? (and who?)  Just because one expert expresses a view, does not prove it’s valid.  Is there general scientific consensus on this theory?  It would also need to be determined whether the beach posting was a result of bacterial pollution or for some other reason.  Just to count postings does not inform the risk.

E.coli Bacteria:

I have viewed E. coli data in the lower river (main stem) going back fifteen years or more.  Generally speaking there is one, two, or no exceedances on any lower river beach over the summer.  Samples had been taken by Water Agency, Regional Board, and SC Public Health Department.  The exceedances were episodic and showed no indication, in my view, of trending septic system break down.  I have a whole file and also need to relocate Water Agency data which is no  longer in printable form on their website.

Staff Report:  Page 4-5: Table 4.1: Some of the locations in Dutch Bill Creek category are nowhere near Dutch Bill Creek, including Unnamed Tributary at River Road near Rio Nido.  Old Monte Rio Rd. appears to come out at Northwood, which is east of Dutch Bill Creek and does not appear to converge with the creek.  Is Diver Dr. a misspell?  I’ve been trying to find locations of these unnamed creeks.  Can you provide a readable map?

Table 4.2 (p. 4-10): Footnote says that locations exceeding water quality objectives are noted in bold font.  I could not see bold font.  Does this mean that all measures were in compliance? Also it appears that the focus was on tributaries as they meet the Russian River.  I guess this is assumed to affect water quality in the RR, yet interestingly the E.coli monitoring data on lower river beaches indicates very few exceedances over the last 9 years.  (I’m referring to water quality data taken by SCWA for the Biological Opinion between 2010 and 2018. See comment above.)   E. coli exceedances noted on page 4-11 at the lower river were located at the mouth of the tributary.  We believe that downstream samples were taken at the next downstream beach.  Data downstream of Dutch Bill Creek would show up at Patterson Point in Villa Grande.  We don’t recall seeing many or any exceedances at that location.   If taken, where, when, and how much showed up?

Enterococci Bacteria:

Table 4.3 on page 4-14 also did not show bold.  Only 6 were in the lower river with 3 of those in upper watersheds.  Three were at the mouth of tributaries (Dutch Bill, Porter Creek, Willow Creek).  Porter Creek is east of Forestville, so it is not in the lower river either.  (Forestville to Jenner)

Bacteroides Bacteria:

In trying to determine source of bacteria, wouldn’t knowing temperature and weather conditions help us know more about the source?  We agree that those HUC-12s with highest median Bacteroides concentrations should have additional monitoring, but E. coli exceedances should really be the determinant.   In the lower river, Green Valley Creek and Porter Creek/Russian River had the highest levels of Bacteroides, with Porter Creek about 3x the level of Green Valley.  Isn’t Porter Creek mainly an agricultural area?  I recall that Bacteroides were more prevalent in high density population areas.  That does not seem to bear out in Table 4.4 (Human specific Bacteroides).  Do you agree?  Why would this be so?  It seems to contradict other statements made about the lower river.  Also, what does all this data tell us about the Bacteroides presence in the various land densities?

Beach Advisories:

Table 4.11 gives the number of advisories for each year but does not say where the postings occurred.  It would be important to have breakdowns by communities (i.e., Forestville, Guerneville, Monte Rio, etc.)  Also it doesn’t say for what they were posted.  Sometimes Public Health puts signs up that are just warnings that say “Swim at your own risk”.  Recently, when there was an illegal wastewater discharge going on, tiny and barely noticeable public health signs were present although signs about the depth and rapid current of the river made swimming conditions dangerous.  I think “Beach Advisories” should be defined here as there are more than one kind.  (See attachment of our back up photos illustrating complaint regarding RRCSD violations.)  Is it possible that sometime beach advisories are not related to water quality but other safety factors.  Also, do warnings about toxic algae count as beach advisories?  Also, is posting the beaches for fecal coliform exceedances adequate for meeting the new standards?  Finally, the regulation is joining what may be two inadequate measures to verify potential danger to public health.  At the same time regulations for most toxic substances go lacking.   Not good!

Table 4-12 gives no indication of how many samples the symbols were based on.  That would be helpful information.

 Chapter 6: Source Analysis

 6.1: Overview

Although not a revised paragraph, the second paragraph is an excellent description of the basis for determining the need for this program.  In other words, it describes the Regional Board’s rationale, as well as indirectly indicates the constraints of controlling all bacteria (or as much as possible) in order to prevent human illness.  It states, “As a matter of general public health protection, there is no allowable load of fecal waste discharge that is without any risk…..An obvious public health principle is to eliminate the discharge of fecal waste to waterbodies that support recreation and other human contact.”  Yet many of us believe that the problem doesn’t merit the heavy handed approach that is planned.  After all, everything we do contains risk.  The food we eat, the cars we drive, the airplanes we fly in, all have risk.  It’s good to mitigate risk, but removing it entirely is probably impossible.

At what burden and what cost to the public? (See Rich Holmer’s letter for suggestions on making this more palatable.)  In all other issues, the Regional Board is called upon to balance the cost of eliminating risk with the potential harm if limitations too burdensome are placed upon the public.  In this case, we see the goal as wanting to remove the entire universe of potentially harmful bacteria (while never considering the potential benefits of nonharmful bacteria) and make Russian River swimming entirely risk free.  While you hold out some carrots, such as a promise to try and get financial assistance for low income septic owners, and an offer of extended compliance time for low income folks who do not intend to expand use of their property, nevertheless, the goal itself does not seem to have been constrained as promised.

Page 6-5 states the following:  “E. coli, enterococci, and Bacteroides bacteria concentrations are statistically the same for wet and dry period runoff draining from developed sewered areas and developed areas with OWTS.  This suggests that certain common features such as leaking septic system/sewer lines may be pervasive.”  This conclusion doesn’t seem to fit when looking at Figure 6.2: E. coli concentrations in wet and dry periods since contributions appear much greater in winter.  While fecal contamination is showing up in the tributaries in summer, there is very little showing up at lower river beaches at that time, even when one considers that swimmers may be adding to the load.  (See graph) We feel that further analysis is called for here.

6.4.7: OWTS

States here: “….parcel density showed a positive correlation with water quality data.  That is, when parcel densities were higher, so were downstream concentrations of FIB.” This is supposedly demonstrated in Figure 6.10:  What was not considered was the presence of large trees on site.  What if high parcel densities had fewer trees and less water was taken up as a result?  Perhaps lower densities had more trees?  There may also be a wealth factor and full time versus part time occupancy that may also figure in to the difference in results with larger parcels having other distinct advantages.   Why weren’t other factors considered?  It seems like conclusions were made based on a very limited number of factors.

Question regarding various river samples designating levels of pollution:  How do you know that the evidence provided for various recreation, septic leaks, and other pathogen sources are all separate pollution indicators instead of a combination of all of the above?

Data for homeless numbers is way out of date: 2005 was used to denote homeless levels on Russian River.  We all know there are far most homeless now than 14 years ago.

Conclusions on page 8-2 don’t seem to match what I read in other sections. (top paragraph) “Non-sewered developed lands are identified as having the highest concentrations of fecal indicator bacteria during the dry season, while sewered developed lands had the highest during the wet season.”  By non-sewered, I assume you mean OWTS served parcels.  My impression of sewered parcels is that they should have no levels during any season unless a lateral is leaking or an SSO is occurring.  Where is data showing E. coli exceedances in lower river in non-sewered areas during summer?  (I know you have bacteroides, but E. coli is the constituent that should really count.)  If you use the phrase ‘fecal indicator bacteria’ to denote bacteroides, and then conclude there is a serious problem with pathogens, that is very misleading.

Page 9-5: I strongly support monitoring discharges from wastewater holding ponds before discharge.  Thank you!

Watershed Monitoring:

One of the issues with all this that comes to mind, is whether the span of time and number of samples were adequate to give a picture of bacterial water quality over time.  Over the years it seems as though the  Regional Board has been extensively under-funded for monitoring.  More recently staff has overcome some of this limitation by asking for more elaborate data from dischargers.  Further, it is anticipated that the R3MP program will make better use of funds available for this purpose.

I’m concerned that in regard to measures to decrease bacterial pollution by homeless encampments and recreators, none of the suggestions in Chapter 9 have much in the way of enforcement power.  The septic system sections have far more detail than any of the other sources.  Is the County to be accountable entity for providing portable bathrooms along the river, keeping them cleaned and maintained, and providing education about and access to their locations?  Who will be responsible? River guides and canoe companies and similar river vacation oriented businesses could form a group to discuss the ways in which a program could be developed and paid for to provide this service.  Perhaps the Sonoma County Water Agency, so proficient in raising funds for their programs, could be persuaded to take the lead in this regard.  After all, the river is the source for everything they do.  Perhaps the resorts should also chip in.  Special events should also chip in.  Historically, people have sat in their canoes during the Jazz Festival, drinking beer and jumping in the river when they need to relieve themselves.

Individual Monitoring & Reporting Requirements:

This section refers to point and nonpoint discharges and states: “Dischargers operating under existing, new or revised NPDES permits or WDRs will be required to monitor, assess, and report on the implementation of their assigned actions, including compliance with the implementation requirements and their effectiveness.”  It is unclear whether this document sees septic systems falling under point or non-point, but the permit holder will be forced to agree with certain maintenance activities (such as assuring in some way that proper treatment, disinfection, and disposal is occurring in a way that does not allow any pathogens to escape into the waterway).   Since most individuals will not be able to do this themselves, it will probably require hiring and expert to do an inspection report.  Is this the 5 year inspection referred to or would this requirement need to occur every month, every year, or what?  The text (page 10-3) states that monitoring wells may be necessary.   What does that involve?

Bacteroides Bacteria:

We will just note that Bacteroides simply indicate the possibility of pathogenic organisms, NOT the actual condition of their presence.  By including bacteroides as part of a triad of metrics (others being E. coli, & enterococcus, actual pathogenic indicators) you will be enforcing, for a condition that may never occur.  Much of the time this document appears to use Bacteroides and E. coli as equivalents.  Did the State intend it that way?

Structural Controls:

This section calls for remedies of the problems of the Russian River County Sanitation District (RRCSD).  It says nothing about whether such remedies are possible.  It does not require that any remedies sought need to fully comply with the TMDL.  It barely acknowledges that numerous attempts occurred over the years to remedy storage and irrigation area issues to no avail.  It does not acknowledge serious problems with RRCSD for current ratepayers, not to mention the problems of adding 1500+ new hookups.  It simply assumes that this is a readily available option.  It alludes to people who have requested including this as a possibility.  We assume that many of those making these suggestions know little or nothing about the problematic history of this system, nor its ability (or inability) to expand.

The statement is made on page 11-43 (bottom) that, “Any potentially significant environmental impacts that cannot be reduced to a less than significant level with mitigation are overridden by project benefits as set forth in the SED.”  We believe that there will be extensive impacts and huge effects if Monte Rio and Villa Grande hook up to RRCSD and that they should not be overridden in any case. Because we are running out of time with this document and because we believe extensive environmental review would be needed for a big expansion of the RRCSD, we will not comment on environmental effects at this time.

As an afterthought however, we add that many, many trees came down in the building of the RRCSD collection system.  This has not been addressed in this document.  The environmental harm could be significant as a result.  (For instance, trees that have their root systems damaged as a result of laying pipes are more susceptible to falling and causing damage, as well as danger for surrounding people, drivers on narrow roads, and properties lining county roads.)

Economic Considerations:

Page 12-6:  Estimate given for storage expansion project planned ten years ago.  No information on how costs have risen in the ten year period and how today’s dollar compares to 2009 dollars.  Also, I seem to recall that Guerneville storage project would only hold 3 mg and not 5.7 mg.   The original EIR projected costs for a 30 mg storage unit because of the need to hold extra water during major floods.  We believe that any new storage facilities need to address the flooding issue.  I believe there is no land close to the plant that could facilitate a much larger pond.

The irrigation project alluded to on p. 12-7 was of a scope that involved designing two 9-12 mile pipelines for a system that produced 250,000 to 400,000 gpd wastewater in summertime.  There was no way our community could use or afford the system as designed.  It was totally unrealistic.  As for irrigation on the Burch property, there was great concern not only adding water in summer to the ancient redwoods, (golf course at Northwood could/would not take more water), but there was an issue of over irrigation at the site and intrusion into downstream ground water.

When you give cost figures for potential future projects, there needs to be some realistic basis for the numbers provided.  The project mentioned were proposed 10-20 years ago.  Also the size of the project would not have been comparable if you would be adding 1500 hookups (Monte Rio and Villa Grande).

We wonder whether estimates are given in 2019 dollars? When you base costs on a ten year old project, you need to translate to today’s dollars.  Also, when talking about technological hardware, it keeps advancing, and each time, the price generally goes up.  So it’s good to compare proverbial apples and oranges.

Finally, continued monitoring of pathogenic bacteria is expected to show improvements in water quality.  Unfortunately, specific causes of improvement are not defined.  Is the monitoring occurring before and after a busy beach holiday, are the tests consistently being taken in the same location under the same circumstances, are there other factors of flow, or temperature, or other conditions that may be affecting the outcome.  None of that has been defined, so no one will know exactly how the improvements occurred.  (page 13-2)

I would be happy to provide any of the documents or information alluded to in these comments.  Please contact me at my email address: rrwpc@comcast.net



Brenda Adelman for RRWPC


PS: I apologize if I have repeated myself here and if sometimes my comments are not well organized.  I often have trouble organizing comments on a 350 (?) page document since I try to go from start to finish.